Rescuing the Lotus Case from its undeserved bad press

Andrew Serdy, 23 July 2026   9 mins read

The 100th anniversary of the collision on the high seas on 2 August 1926 between the French and Turkish steamships Lotus and Boz-Kourt, sinking the latter with the loss of eight lives, is imminent. It offers a suitable occasion to ponder why the judgment of the Permanent Court of International Justice (PCIJ) in the ensuing case between  France and Turkey nowadays meets near-universal disapproval.

The Court held that Turkey’s institution of criminal proceedings against Lieutenant Demons, officer of the watch on the Lotus at the time of the collision, did not conflict with the international law rules on jurisdiction. Lotus seems second in notoriety only to the South-West Africa Cases for which the received view is that the International Court of Justice (ICJ, the PCIJ’s successor) wanted to make amends by introducing the concept of obligations erga omnes in the Barcelona Traction Case only four years later, that if applied in South-West Africa would have produced the opposite result. Admittedly Lotus has not been so starkly disowned. Rather the closest the ICJ has come to that is the observation in the Arrest Warrant Case that the principle it represents has reached its high-water mark and will not be expanded further.

Even so, it often feels close to being socially unacceptable among international lawyers to support the Lotus case as correctly decided, which is nevertheless what this post will do. This is despite the fact that the international community acted to reverse the result of Lotus subsequently:

All three instruments provide that in cases of collision or any other incident of navigation on the high seas, penal jurisdiction over the acts causing the collision or incident is exercisable only by the flag State of the ship on which the accused is serving, or the accused’s own State of nationality.

A defence of Lotus is somewhat less disreputable now thanks to the belated – if only implied – confirmation of the Lotus reasoning in a case with roughly parallel facts, The “Enrica Lexie” Incident (Italy v. India) decided in 2020 by an arbitral tribunal formed under UNCLOS. Here UNCLOS Article 97(1) was held inapplicable to the fatal shooting of two Indian fishermen mistaken for pirates by a security detachment of Italy’s armed forces aboard an Italian merchant ship, 20 miles off India’s coast and thus in its exclusive economic zone, assimilated to the high seas for navigational purposes by Article 58(2). This was because the fact that the detachment played no role in setting the ship’s course and speed meant that the incident was not one “of navigation”. Hence India – parallelling Turkey in Lotus – was found to have jurisdiction concurrent with Italy’s over the incident as a whole. For a more detailed exposition of the arguments, see this chapter by Vrancken in a recent volume co-edited by the author.

Much of the criticism directed at Lotus, however, is by commentators taking issue with propositions that the majority judgment does not in fact contain or endorse. The misattribution tends to happen in blogposts rather than journal articles, perhaps a positive sign that peer review is serving its intended purpose.

Five such propositions are treated below, but first it is helpful to disentangle the Lotus “principle” – that in international law anything that is not prohibited is permitted – from the narrower Lotus jurisdictional “rule”: when ships collide the flag States of both ships have jurisdiction to try any offences alleged to have been committed in causing the collision, provided at least one element needing to be proved in order to secure a conviction occurred aboard the ship flagged to the prosecuting State. The principle attracts more of the opprobrium than the rule. Yet it is not necessary to embrace it in order to say that the case – based on the narrower rule – was correctly decided. Both principle and rule have given rise to dubious propositions.

The first two propositions concern the principle and bespeak an impatience with positivism, which has rather gone out of fashion, replaced by the notion that States are dangerous entities that must be constrained as far as possible, not left to do largely as they please. The present author has no particular philosophical objection to positivism, but has no need to take a firm position on it beyond noting that, in his experience as a former government official, States themselves, who after all make the international law by which they are bound, are positivists 98% of the time. The remaining 2% represents occasions when litigating States – invariably applicants, never respondents except when counterclaiming? – try to persuade an international court or tribunal that the respondent has some obligation it is alleged to have breached despite never having expressly consented to it.

Absurd proposition 1: Taken to its logical conclusion, the Lotus reasoning makes universal jurisdiction the default rule.

Nobody actually believes that universal jurisdiction is the default rule, not even Belgium despite in effect making that argument to defend its universal jurisdiction law at issue in Arrest Warrant, there being no prohibition as such of universal jurisdiction. The source of confusion here is that, as presented in modern textbooks, the international law rules on jurisdiction solidified in the century since Lotus are framed positively, rather than as a double negative (i.e. “nothing prevents States from…”): States have jurisdiction over events and persons within their territory and over events and persons outside it in a limited number of other situations. Among the latter is universal jurisdiction, but it is a narrowly scoped exception. This is incompatible with the everything-permitted-unless-prohibited caricature, which would force a start from a position of complete freedom.   

Absurd proposition 2: Absence of a prohibition of X is a positive right to do X and shields against incurring responsibility for incidental breach of some other obligation associated with X.

Confusing positive permission with double-negative non-prohibition is also at the root of this error, into which it is all too easy to slip because the result in any given instance will be the same either way: the conduct in question can lawfully be engaged in, either because it is expressly allowed or because there is no rule against it. This can lead dangerously astray those who yield to the temptation to take an extra step: a positive right to do X elides into an absolute right to do it, rendering irrelevant all other potentially applicable law, such that a prohibition against Y can be ignored when Y happens as a result of doing X. Lotus offers no support whatever for that reasoning, and should not be blamed for it.

The other three propositions rest on the Lotus jurisdictional ruling and stem from various misreadings of the majority judgment.

Absurd proposition 3: The Lotus decision endorses the passive personality basis of jurisdiction.

This proposition was formerly commonly seen but less so recently. True, passive personality was the basis on which, under the Turkish Criminal Code, the Istanbul authorities prosecuted Lieutenant Demons. Judge Moore, dissenting, deprecated it, but Turkey itself did not advance passive personality as a defence before the PCIJ. Instead, it argued, and the Court accepted, that the fact that the loss of life occurred on board a Turkish-flagged ship gave Turkey concurrent quasi-territorial jurisdiction over the offence of manslaughter, which would not have been completed had nobody died. For jurisdiction is rarely exclusive. Turkey had no need to, and did not, maintain that France lacked jurisdiction, only that Turkey too was entitled to exercise it. This leads to

Absurd proposition 4: Ships are floating territories of their flag States.

It is true that Judge Nyholm, dissenting, stated as much, and Judge Loder too by implication appeared to agree, but this is not the approach taken by the majority judgment. What the Court actually said was more subtle: it merely assimilated ships to territory for jurisdictional purposes, so avoiding a legal vacuum on the high seas. UNCLOS Article 92 comports with this. Under it, a ship on the high seas is a place where the flag State’s law applies, but that falls short of asserting the ship is its territory. Thus a collision between ships of different flags brings two such quasi-territories into momentary adjacency, akin to neighbouring States on land where an act in one produces effects in the other and concurrent jurisdiction is not controversial, and an on-water shooting is little different. (Imagine the body of water over which fatal gunshots were fired in Enrica Lexie not as the ocean but as a border river.)

Absurd proposition 5: By endorsing Turkey’s jurisdiction over an event occurring on the high seas, the Lotus decision promotes extraterritoriality, a legal evil.

This overdramatises the problem of extraterritoriality. Though sometimes undesirable, it is no evil, just an unavoidable consequence of the way territorial sovereignty over the dry parts of the Earth’s surface is divided among States. There is no intrinsic problem with concurrent jurisdiction, even if for commercial certainty it is common to eliminate one or more possibilities by virtue, for example, of a choice of law and choice of forum clause in cross-border sales contracts. The issue is a practical one, of keeping extraterritoriality within reasonable bounds, so that, merely by acting quickly, a State with a weak nexus with the events in question cannot too easily impose and enforce its own law in relation to them, displacing the law and courts of a State with a stronger connection that appears disadvantageous to the party seeking to circumvent it. Unilateral sanctions, if applied in this controversial way, are rightly faulted.

So, concluding, the passage of 100 years has changed the context in which we should think of the Lotus case. International law has become significantly more sophisticated and complex and the Lotus principle no longer gives, if it ever did, carte blanche to States to have their head. There are, in other words, far more prohibitions to which they are subject, indirectly as well as directly, than there were in 1927. While double negatives are commonly criticised on stylistic grounds, let us thank the Lotus case, both principle and rule, for its salutary reminder that in law it is often a mistake to substitute them with a positive.

Adapted from the slides accompanying a presentation entitled “Rescuing the Lotus Case from its undeserved bad press: Absurd propositions attributed to the case for which it does not stand” delivered at the 8th online seminar of the Lotus 100 Project, 1 December 2025.

Free Elections and “Turning Down the Noise” on Online Platforms

Michael John-Hopkins, 17 July 2026      11 mins read

In July 2025 in Bradshaw and Others v United Kingdom, the European Court of Human Rights confronted the unexpected tension between freedom of expression and free elections. Democratic societies must preserve the broad freedom of political expression on which open electoral debate depends, but that freedom may also be exploited through coordinated foreign influence operations designed to undermine free elections.

The applicants argued that the United Kingdom had failed to adequately investigate, and protect against, Russian interference in its democratic processes. The Court did not find a violation. It emphasised the difficulty of measuring the effects of disinformation and influence campaigns and allowed the United Kingdom a wide margin of appreciation, particularly because no clear European consensus has yet emerged on the specific protective measures States must adopt. The judgment was nevertheless significant. It accepted that Article 3 of Protocol No 1 of the European Convention on Human Rights (right to free elections) may impose positive obligations where hostile-State interference creates a real risk of impairing the very essence of the right to free elections.

Before the judgment, the UK Parliament Intelligence and Security Committee’s Russia report (2020) had already described the United Kingdom as a target for Russian disinformation and political influence operations. Importantly, it distinguished disinformation from the broader playbook of malign interference, which may combine manipulated narratives with illicit funding, cyber operations, interference with electoral machinery, covert amplification and “hack-and-leak” activity. Such operations frequently seek less to secure agreement with a single message than to intensify division, confuse attribution, discredit reliable sources and erode trust in democratic institutions.

In Bradshaw, the Court accepted that there was evidence of interference in the United Kingdom’s democratic processes of sufficient intensity to be capable of impairing the very essence of the right to benefit from free elections. Its conclusion was not that the threat was unreal, but that the United Kingdom’s response remained within its wide margin of appreciation.

That tension captures one of the central problems now facing democratic legal systems. Electoral rights can no longer solely be guaranteed at the ballot box but must also be protected in the communicative environment in which electoral preferences are formed. That environment is increasingly structured by platforms, recommender systems, political advertising, synthetic media, data profiling, covert funding and the speed with which false or manipulated material can outrun scrutiny and correction.

The legal challenge is therefore not simply to ensure that voters can physically cast a ballot. It is to protect an information environment that remains sufficiently open, plural, intelligible and contestable to ensure deliberative democracy – without converting the protection of elections into a general power to police political truth.

“Turning down the noise” in the digital environment

This is where the new radical idea of “turning down the noise” or “protection from noise” is useful. The phrase should be handled carefully. It does not describe a right to truth, silence, comfort or protection from disagreement or offence. Democratic politics is “noisy” by design. It includes exaggeration, error, satire, anger, mobilisation and sharp contestation.

The concern is narrower, namely organised, synthetic and amplified noise. That means engineered saturation, impersonation, microtargeting, synthetic deception, covert amplification, harassment or suppression of correction in ways that materially impair the practical ability to receive, test and contest political information autonomously.

In his separate concurring opinion in Bradshaw, Judge Jakab gives this idea one of its first significant judicial formulations within the Strasbourg system. He suggests that digital disinformation can reach a point where the “noise” level is so high that the right to receive information under Article 10 may itself be impaired. His concurring opinion fuses and balances human rights, platform governance, and national security considerations, at both high and granular levels of analysis.

One of the strengths of liberal democracies when compared to autocracies is that liberal democracies’ political decisions and democratic accountability mechanisms (such as elections) tend to be relatively more strongly influenced by fact-based discourses weighing arguments. If we allow the digital disinformation noise levels to ravage as high as they currently are in our public debates, then we are endangering liberal democracies’ structural superiority.

This may become an important line of jurisprudential development because it sits precisely where law, politics and technology now meet. There is also a historical reversal here. The post-1989 European human-rights order was significantly concerned with supporting transitions towards democracy and the rule of law. It must now consider how established democracies should be protected when revanchist authoritarian states and resurgent anti-democratic movements exploit democratic openness, sometimes in mutually reinforcing ways, to destabilise political institutions and corrode public trust from within. It must also confront the reflexive danger that the language of democratic protection may itself be exploited. Real, exaggerated, fabricated or unverified claims of malign interference may be weaponised to delegitimise elections, discredit political opponents and independent institutions, justify coercive or exclusionary measures, and consolidate power.

Giving citizens more control over the noise

The point is not that Strasbourg should prescribe a single model of platform regulation. Bradshaw itself points the other way in that in an area of technical complexity, national-security sensitivity and limited European consensus, States retain a wide margin of appreciation. But subsidiarity does not mean that human-rights law has nothing to say. As Robert Spano has argued in a different context, the contemporary Convention system increasingly works through process-based supervisory review whereby national authorities are given space where they act in good faith, reason carefully and balance competing interests, but law still helps sustain the democratic character of political decision-making. The question for future cases is whether emerging regulatory practices concerning provenance, amplification, synthetic content, political advertising and platform risk assessment will give courts firmer ground on which to assess the informational conditions of democratic choice, particularly where those conditions are deliberately manipulated to undermine the orientation towards “effective political democracy” identified in the Preamble to the European Convention of Human Rights.  

In practical terms, protection from noise would not mean that States or platforms decide political truth for citizens. It would mean giving citizens more meaningful control over the conditions in which political information reaches them. Users should be able to know whether they are dealing with a real person, a pseudonymous account, a bot, a political advertiser, an influencer, a campaign organisation or a verified media source. They should be able to reduce the volume of automated, recommended or paid-for material; choose non-personalised or following-only feeds; see when political content is synthetic or AI-generated; and understand why they are being targeted. The aim is not to remove disagreement from democratic life, but to make the architecture of online political communication more transparent, adjustable and contestable.

2025 Romanian presidential election – noise control or authoritarian regression?

The Romanian election litigation shows the stakes. In Călin Georgescu v Romania, the Strasbourg Court rejected the application concerning the annulment of Romania’s 2024 presidential electoral process, principally because Article 3 of Protocol No 1 does not normally apply to presidential elections unless the office has sufficient legislative functions. That means Georgescu should not be treated as a Strasbourg merits ruling approving the annulment. Its importance is different. It records the kind of factual scenario now pressing against electoral law, namely algorithmic amplification, digital technologies and artificial intelligence, unlabelled electoral promotion, campaign-finance opacity and unequal campaign visibility.

The Romanian Constitutional Court’s Ruling No 32 of 6 December 2024 treated the electoral process as vitiated by non-transparent digital promotion, misuse of social-media algorithms, artificial intelligence, and undeclared campaign funding. Whether one sees that decision as necessary constitutional safeguarding or a dangerous authoritarian regression, it illustrates the outer edge of protection from noise. At some point, information-environment manipulation may be argued to affect not merely campaign fairness, but the free formation and expression of democratic choice.

That is why the Venice Commission’s urgent report on the cancellation of election results is so important. It recognises that online campaigning, disinformation, artificial intelligence and external influence create new electoral risks. But it also insists on constitutional discipline. Generally, annulment must be exceptional, it must rest on clear legal competence, fair procedure, adequate evidence, and a reasoned assessment that irregularities may have affected the result. In other words, democratic systems must be able to recognise information-environment harm without making election cancellation too easy.

The emerging regulatory picture is therefore fragmented but significant. Bradshaw gives the human-rights law ‘seeding’ that electoral rights may require positive protection against hostile interference. The Romanian litigation gives the factual pressure point, namely platform amplification and digital opacity can put electoral legitimacy under severe strain. The Venice Commission supplies the rule-of-law brake in policy terms that serious remedies require legality, evidence, procedure and proportionality.

Retreat from platform self-governance in the EU and UK

Platform governance is the next part of the picture. The EU Digital Services Act moves beyond a purely passive view of platforms. It treats very large online platforms and search engines as potential sources of systemic risk, including risks to fundamental rights, civic discourse, public security and electoral processes. The Commission’s electoral-process guidelines under the DSA translate that idea into election-specific concerns including recommender systems, generative AI, political advertising, coordinated manipulation, crisis-response mechanisms and cooperation with electoral authorities.

The EU Regulation on the transparency and targeting of political advertising adds another layer. Political advertising becomes democratically dangerous when voters cannot tell who paid for a message, why they are seeing it, which audience was targeted, or whether different voters are receiving incompatible versions of a campaign. Transparency does not solve manipulation, but it helps restore the public character of political contestation.

Privacy and data protection also matter. Profiling and political microtargeting can transform democratic communication into the behavioural sorting of citizens and the tailored manipulation of their choices. The issue is not merely privacy in the narrow sense. It is whether voters are placed into opaque political environments selected according to inferred fear, resentment, identity or susceptibility. In that respect, the GDPR is also part of the democratic-resilience architecture of mitigating manipulation and coercion.

The EU AI Act is especially relevant because it recognises that harm may arise from system design, not only unlawful content. Its rules on manipulative or deceptive techniques, vulnerability exploitation, synthetic-content transparency and AI systems used to influence elections or voting behaviour do not create a general law of malign influence. But they identify risk indicators that map closely onto protection from noise, namely opacity, synthetic realism, behavioural steering, vulnerability targeting and loss of provenance.

In contrast, the UK relies on a dispersed framework comprising the Online Safety Act 2023, data-protection law, electoral law, including campaign-source transparency requirements under the Elections Act 2022, and the National Security Act 2023. The latter criminalises certain coercive, deceptive or otherwise unlawful foreign interference with political processes, public functions and UK interests, while the Foreign Influence Registration Scheme requires greater transparency concerning political influence conducted at the direction of foreign powers.

European regulatory leadership of noise control in the digital sphere

This is where a limited ‘Brussels effect’ may matter. The point is not that the EU model will become global law, nor that it is beyond criticism. But large platforms, AI providers and advertising intermediaries may find it easier to build systems around demanding expectations of transparency, provenance, audit, risk assessment and researcher access. If so, EU regulation may indirectly raise the baseline for democratic information governance elsewhere, including within the broader European legal space. That hope depends on enforcement, judicial oversight, platform incentives and whether other jurisdictions view systemic-risk regulation as democratic infrastructure rather than censorship.

The importance of Bradshaw is therefore not that Strasbourg has created a settled right to protection from disinformation. The Court recognised the possible relevance of positive obligations but, rather than prescribing particular safeguards or applying strict review, left States considerable discretion within a wide margin of appreciation. That caution reflects the absence of a clear European consensus on the measures required to protect elections from hostile information operations. It also makes it more important to identify where regulatory convergence, and potentially such a consensus, is beginning to emerge. Across human-rights law, electoral law, platform regulation, data protection, AI governance and Venice Commission standards, recurring principles include provenance, transparency, recommender accountability, limits on microtargeting, synthetic-media labelling, campaign-finance visibility, platform-data access, fair procedure and evidence-based remedies.

None of this means that technology can solve social, political, economic and diplomatic problems. Democracies cannot outsource electoral integrity to labels, dashboards, audits or content-moderation tools. But the opposite error is equally dangerous, namely leaving platforms, AI systems and advertising infrastructures without legal guidance, free to operate in a manufactured aterritorial space, immune from accountability for the destabilizing effects occurring through their digital architectures while expecting voters, journalists, candidates and electoral bodies to absorb the forced choices and consequences.

An emerging consensus on noise control?

The emerging consensus, if there is one, is not a consensus on truth control. Nor is it yet a fully settled Convention-law consensus on exactly what every State must do. There remains regulatory divergence, and much will depend on domestic implementation, enforcement capacity and institutional design. But the legal and policy conditions for convergence are now visible. Across platform regulation, electoral law, data protection, AI governance, political-advertising transparency and Venice Commission standards, recurring concerns can be identified, namely provenance, transparency, recommender accountability, microtargeting limits, synthetic-media labelling, campaign-finance visibility, platform data access, fair procedure and evidence-based remedies.

Protection from noise is best understood in that restrained sense. It is not protection from disagreement, nor a licence for State truth control. It is protection of the practical conditions that allow disagreement to remain visible, intelligible, contestable and democratic. Citizens should be able to know who is speaking, who is paying, whether content is synthetic, why they are being targeted, how platforms amplify political material, and what remedies exist when those systems are abused. The deeper danger is not only that citizens may encounter false claims, but that digital architectures may increasingly shape attention, visibility and behaviour in ways that are difficult for citizens to perceive, contest or resist. In that respect, Shoshana Zuboff’s account of “instrumentarian power” is useful. Malign actors, foreign, domestic, may exploit these  aterritorial architectures not by physically silencing or controlling citizens through totalitarian force, but by steering the conditions under which they see, feel, rank, trust and act in a way that is detrimental at individual and societal levels. It is to preserve enough autonomy, transparency and contestability for us to be more confident that what we take ourselves to think is genuinely our own.

Is a Social Media Ban for Under 16s a Panacea for Online Child Safety?

Stephanie Law, 3 July 2026       8 mins read

On Monday the 15th of June 2026, a week before announcing his resignation as Prime Minister and leader of the Labour Party, Keir Starmer confirmed that the Government would introduce legislation to implement a social media ban for children and young people in the UK.

The under-16s social media ban reflects a significant step in the UK’s attempts to regulate the online environment, and more specifically, to provide for “a safer digital childhood”. Indeed, for the supporters of a blanket age-based ban, the Labour Government’s commitment constitutes a necessary intervention to protect children’s safety and wellbeing, amongst growing public concern about social media use and the risks arising from children’s access to social media content that may exacerbate mental health concerns and existing vulnerabilities, expanded opportunities for engagement amounting to cyberbullying or exploitation, exposure to material relating to self-harm, suicide and eating disorders, as well as hate speech and misinformation.

While the background of the broader discussions on children’s online safety encompass many tragic incidents, it is necessary to consider whether young people’s exposure to risk in their increasingly intertwined offline and digital lives will effectively be reduced by an age-based ban. Until now, the UK has stopped short of introducing a blanket ban on young people accessing social media platforms. Instead, the Online Safety Act 2023 establishes obligations on platforms to remove illegal content, and content which is harmful to children.

The popularity of social media bans 

The UK’s proposal follows a similar ban implemented in Australia at the end of 2025 while proposals have also been advanced in France, Portugal and Spain. At the EU level, the European Parliament adopted a resolution in November 2025 proposing that the EU adopt a minimum age requirement of 16 for access to social media and video sharing platforms, as well as AI companions. Young people between the ages of 13 and 16 would be able to access such platforms only with parental consent. This age-based ban is envisaged to operate alongside and strengthen the enforcement of the EU’s Digital Services Act (similar to the UK’s proposed ban).

The UK proposes to ban under 16s from accessing user-to-user social media platforms, that is to say, platforms that allow for social interactions between users and on which content can be posted; this would include platforms like Facebook, Instagram, X, Snapchat and TikTok. Moreover, the Government intends to implement a requirement that access to certain harmful features on platforms, such as livestreaming, stranger communication (including in gaming) and “romantic companion” chatbots, is restricted to over 18s.

The ban will not apply to direct messaging platforms, like WhatsApp. With these measures, the Government hopes to adopt a model for protecting children online, which goes beyond a simple blanket ban, and instead responds to both the risks children face and their broader experiences online. These risks, the Government argues, are exacerbated by the real-time accessibility of potentially harmful content and tech companies’ use of algorithmic recommender systems, which undermine the scope for effective content moderation.

The decision to introduce a ban follows a short consultation, undertaken between the months of March and May of 2026, in which the Government sought the views of the general public, including parents, guardians and young people, children’s charities and civil society organisations, as well as other stakeholders, including teachers, academics and industry partners. The consultation received over 115,000 responses, with the Government concluding that 9 out of 10 parents and two-thirds of children support restrictions on under-16s accessing certain social media platforms.

The timeframe set out by Keir Starmer and Liz Kendall, the Secretary of State for Science, Innovation and Technology, reflects the Government’s prioritisation of the social media ban. Starmer indicated that the Government would “move at speed” to drive secondary legislation through Parliament using the Children’s Wellbeing and Schools Act 2026, with the intention of putting the ban into place in the Spring of 2027.

The Government has indicated that Ofcom will be responsible for establishing “accurate, robust, reliable and fair” age assurance mechanisms which protect privacy rights, with the aim of ensuring that children cannot work around these restrictions. Platforms will be responsible for implementing these mechanisms, meaning that the platforms and not children themselves will be penalised for non-compliance. Ofcom is designated as the independent regulatory body for online safety in the Online Safety Act 2023; it is competent to identify – through the adoption of guidance and codes of practice – the practices that online platforms should follow to ensure compliance with their obligations, monitor, investigate and assess compliance and take enforcement measures.

The problems with a blanket age-based ban and its alternatives

The decision to restrict under-16s access to social media platforms is by no means uncontroversial. While there is broad consensus across law and policy-makers, public health researchers and child protection advocates that children and young people face a number of risks from accessing specific types of content and from various forms of online engagement, the nature and severity of the risk of those harms is not homogenous. This necessarily gives rise to questions as to the appropriateness of a blanket age-based restriction, relative to alternative interventions that would still allow children and young people to benefit from the opportunities that digital technologies provide. These include – amongst others – access to educational resources (indeed, the UK Government has indicated education platforms are not targeted by the ban), the development of socialisation skills, social inclusion and integration into supportive communities, and exposure to diverse cultures and viewpoints.

The problematics of the ban are significant. There is a danger that the implementation of a blanket ban merely gives the appearance of a Government being proactive, creating a false sense of security as to being in control of a significant problem, in a way that may undermine scope for further regulatory intervention. Regulating a rapidly evolving and complex digital environment driven by a small number of powerful tech companies necessitates informed and critical discussions about the role of technology in society.

While the UK Government indicates that online platforms will be required to implement effective age assurance mechanisms, there is a possibility that parents, guardians and educators will bear the burden of “policing” children’s access and use of technologies that are constantly developing.

As research on the Australian experience illustrates, 85% of under 16s continue to use social media platforms, either by bypassing “simple” age verification mechanisms (i.e. being asked to enter their age), using fake profiles or (a challenge identified by the UK Government), on more rare occasions, using a VPN. Six months after its implementation, the effectiveness of the ban is not yet clear, however the Australian Government has since enacted new legislation to empower its regulator to monitor, enforce and punish non-compliance with the ban.

The UK’s under-16s social media ban must be executed with a view to also implementing alternative interventions that would target the effective enforcement of content moderation obligations by Ofcom. At the same time, a safe digital environment requires safer platform design, and access to resources that aim at improving digital literacy (with enhanced provisional support for the most vulnerable of young people).

Platforms as the real culprits?

On the one hand, the way in which online platforms are designed enhances the risk of harm to children by maximising user engagement. Such design choices include the use of addictive design features and recommender systems, weak default privacy settings and limited transparency, behavioural profiling and targeted and personalised advertising.

Research suggests that digital markets are characterised by substantial power, control and knowledge asymmetries between individuals and tech companies, which leads to an erosion of autonomy, manipulation, exploitation, and addiction as well as broader societal harms to fundamental rights and democratic participation. Digital environments are characterised by digital asymmetry, which describes the structural and systematic conditions of these systems, and leads to digital vulnerability, which is dynamic, contextual and structural; every individual is potentially vulnerable, a condition which may shift depending on the digital environment with which they are faced.

Children and young people may be particularly vulnerable (and are treated as such, for example, in consumer law), with social media exacerbating existing vulnerabilities (through for example, the use of recommender systems which direct vulnerable individuals to increasingly extreme content).

Digital literacy to prepare children for their digital adulthoods

On the other hand, greater digital literacy is required; young people under 16 will, within a few short years, be faced with digital environments which will very likely be vastly different, more complex and more dangerous than those currently in existence. In order to be able to navigate these regimes safely, young people must be equipped with the knowledge and critical analytical skills to – inter alia – recognise misinformation and malicious communication, assess algorithmic recommender systems and manipulative design features, make informed decisions and respond effectively to the risks arising from harmful online content. Under the Online Safety Act 2023, Ofcom has a key role to play in promoting media literacy.

However, what is clear is that the question as to whether the UK’s proposed social media ban for under 16s is appropriate and proportionate cannot be framed as giving rise to a binary choice; restrictions on children and young people’s access to social media must be implemented hand-in-hand with effective content moderation, the regulation of platform design and digital literacy endeavours.

Corporate deregulation – time for investors to play their voting cards?

Brenda Hannigan, 1 June 2026       7 mins read

In an ill-advised move, the board of BP recently rejected a shareholder resolution so preventing its circulation to shareholders ahead of the company’s annual general meeting (AGM). As the FT reported (11 March 2026), the resolution addressed strategies to maintain shareholder value in the not unlikely event of a decline in demand for fossil fuels. It was filed by a group of 16 institutional investors and some retail investors brought together by a Dutch green investor group, Follow This. 

The board’s rejection of the resolution made the news because we are accustomed to corporate boards maintaining at least a semblance of co-existence alongside shareholder engagement, empowerment and stewardship. Not in this case. In all probability, the resolution would have gone almost unnoticed by the media, had it not been rejected by the board. 

The episode seemed to reflect in many ways the current position. There is a business clamour for capital, flexible markets and a wider pool of investors. The Government has responded with deregulation initiatives aimed at driving ‘growth’ and ‘competitiveness’. The consequence, whether or not intended, is an increase in board autonomy and a consequential diminishing of the role of shareholders. The cumulative effect of recent changes should be a wakeup call for shareholders who need to find their voice and show a greater willingness to challenge boards.

Regulatory retreat by the Financial Conduct Authority

The deregulation agenda is driven, in part, by the collapse in London listings over the past decades (for many reasons including the growth of private capital and the perceived regulatory burden of being listed). In March 2009, there were 1130 UK listed companies, by March 2026, there are 764. It is a similar picture on AIM, the Alternative Investment Market. Hence there have been numerous reviews of listing and capital raising (Hill (2021), Kalifa (2021), Austin (2022)) , followed by a new UK listing regime (UKLR) . It was devised by the Financial Conduct Authority (FCA) with the intention of attracting listings and investors to a deregulated market and came into effect in July 2024. 

A significant element of the new listings regime is the removal – despite robust investor objections – of shareholder approval of substantial transactions and related-party transactions (i.e. with directors, substantial shareholders and associated companies so potentially rife with conflicts of interest). The requirement for approval was long regarded as a defining component of the gold standard, London premium, listing. Following consultations, the FCA was persuaded by strong support for abolition (unsurprisingly) from the companies themselves, their legal advisers and UK market operators [paras 6.16, 6.19, 6.29]. 

The FCA noted that companies would not proceed in any event with such transactions without the support of their largest shareholders – in effect, acknowledging that only shareholders in the inner circle matter [para 6.50 to 6.57]. It considered that shareholders have other tools such as statutory shareholder meetings (invariably post hoc) and the ability to requisition meetings (only if they hold more than 5%, and unlikely in practice). Of course, the FCA also noted, shareholders can remove directors, but this nuclear option is rarely publicly exercised, or they can disinvest. The latter option seems a strange suggestion given part of the policy background here is precisely to stop disinvestment and attract investors to London listings. 

Legislative retreat by the Department for Business and Trade

If the FCA is the inner ring of this deregulation, the outer ring is the Department for Business and Trade (DBT) which has taken an increasingly deregulatory stance.

In January 2026, the DBT, under a Labour Government, abandoned the long-promised and much-anticipated Audit Reform and Corporate Governance Bill — legislation that had its origins in the Brydon Review (2019) and which was intended to increase oversight and accountability, not just of corporate auditors, but also of directors. Investors would have benefitted potentially from these reforms, given they struggle with oversight of the directors. The ‘difficult decision’ (in the words of the DBT Minister), after six years of work, was hidden in a lengthy DBT press release.

In this regard, the DBT was following in the footsteps of the previous Conservative Government which, in October 2023, reneged on a commitment, again reflecting Brydon, to introduce requirements for a variety of board statements on matters such as distributable profits, distributions, risk management and resilience, including the risk of material fraud. Such statements would have forced boards to verify and stand by this information. In this case, the withdrawal came even though the draft regulations were already before Parliament. 

These U-turns would suggest effective lobbying from big business and their lawyers, under cover of the many economic and political crises of current times, to reduce transparency and block any measures which might expose directors to any individual accountability. The result, if not the declared aim, of these changes is to reinforce the power of corporate boards and weaken shareholder democracy. 

The withdrawal of the ‘public register’ of shareholder dissent 

A further example of successful lobbying is the DBT’s announcement, in November 2025, that the Government had asked the Investment Association to cease to maintain the public register. This public register was set up in 2017 and recorded significant shareholder dissent, categorised for this purpose, in accordance with the UK Corporate Governance Code (2024), Provision 4, as votes against the board recommendation on a resolution of 20% or more. 

The request is inexplicable: why ask the Investment Association to discontinue a mechanism which usefully collected the relevant information in one place? The information is publicly available in any event but would require a searcher to have the patience to trawl through the National Storage Mechanism, company by company. There is no cost to the companies or to the Government. The only explanation is that the DBT bowed to business pressure: boards disliked being featured on a register which alerted investors and other interested stakeholders to shareholder discontent, especially if the company appeared (as some did) year after year. 

Ironically, BP is one of the beneficiaries of this deregulatory change. Proving that shareholders can resist board demands, BP shareholders did vote down two resolutions supported by the board including refusing to accept a new set of articles (52.88% against) which, inter alia, would have allowed the directors to hold only virtual (electronic) AGMs. They also voted heavily (74.15%) for a shareholder resolution calling for disclosures on exploration capital expenditure and the creation of shareholder value which had been opposed by the board. Had the public register of dissent still been in use, BP would have been an entry on it. 

More deregulation to come – what shareholders must do  

There are other deregulation initiatives in the pipeline. DBT plans to allow virtual AGMs (vigorously opposed by institutional shareholders as a lesser form of public accountability, giving boards greater control of shareholder engagement than a physical meeting) and intends to streamline (i.e. reduce) corporate reporting, primarily for the largest companies. Of course, a switch to virtual meetings will need shareholder approval since commonly it will require an alteration of the company’s articles of association. Shareholders will want to think about whether they are willing to support such a change and the implications for their engagements with the company. 

The DBT also intends to deregulate (reduce) disclosure by the largest companies so investors need to reflect on what information they are willing to do without and be prepared to stand their ground on reduced disclosures if they are unhappy. Let the largest institutional shareholders vocally oppose related-party transactions, even if they no longer have a vote on the matter. 

Contrary to initial expectations, it may be that the result of the deregulatory measures favouring boards is to force a level of engagement and voting from shareholders which has not previously been the case. Shareholders need to realise that they do have a weapon – their votes – and they should make considered use of them.

The Killing of a Murderer – a Rebalancing of the Order of Things?

David Gurnham, 15 April 2026     7 mins read

On 7 March 2026, I caught a BBC television news headline reporting that the murderer and former school caretaker Ian Huntley had died in prison following severe head trauma inflicted by a fellow prisoner at HMP Frankland, Country Durham, in an attack a week or so previously. The newsreader having given this much of the story, I expected the report thus introduced to offer more about the incident itself: at the very least some details about the man suspected of killing Huntley and facts so far established about the circumstances and conditions at the prison.

But the report (by the BBC’s Senior UK Correspondent Seema Kotecha) was not like that at all. Her full story, taking precisely two minutes of airtime, ran thus:

Soham in Cambridgeshire, an ordinary Sunday in August. Jessica Chapman and Holly Wells were best friends. Mad about Manchester United, this photo was taken by Holly’s Mum that afternoon. At around six, they left Holly’s home, it’s thought to buy sweets. The search lasted 13 days; more than 400 officers took part. Amid huge public concern, it became one of the biggest operations of its kind. A school caretaker, Ian Huntley, said he had seen the girls before they disappeared.

His girlfriend Maxine Carr was Holly and Jessica’s teaching assistant. The girls’ bodies were finally found by a gamekeeper, dumped in a ditch around 10 miles from Soham. Huntley was arrested the same day. He’d lured Holly and Jessica into his home, claiming his girlfriend was inside. He was sentenced to life with a minimum of 40 years, the judge said his continuing lies and manipulation had deepened the family’s suffering.

Maxine Carr, who’d given him a false alibi, was sentenced to three and a half years for conspiring to pervert the course of justice. Huntley, as one of Britain’s most notorious prisoners, was always a target. He was slashed across the throat in 2010, needing 21 stiches. The man who senselessly stole the futures of two happy, carefree girls enjoying their summer holidays.

I found it odd that the BBC report seemed to be so reluctant to talk about Huntley’s own death, a reluctance apparently reflected the general attitude of people of Soham as reported in another BBC story from the same day: that ‘he’s not worth their breath or time’.

Having listened to the report, my attention must have been taken away to something else, because it was only when coming back to it a month later that I found that it had been followed up with an interview with Danny Shaw (former BBC journalist and crime and policing commentator) who had elaborated on some of the facts of the instant case that I had expected to hear in the initial report.

Shaw also mentioned certain contextual matters: that the same prison had seen a serious attack on prison staff by Hashem Abedi in April 2025, and that there had been a sharp rise in the number of prisoner deaths by homicide in the UK in 2025 and 2024 (as reported by the Ministry of Justice). He also noted that the principle of rehabilitation is undermined in the case of prisoners serving a whole life term (as Huntley’s alleged killer was) since they have ‘nothing to lose’ by further offending.

But the first minutes of his contribution, too, continued in the same vein as Kotecha’s, reminding viewers of the ‘shocking’, ‘brutal’ and ‘evil’ nature of Huntley’s own crimes and his attempts to ‘manipulate’ both the girls, and afterwards, the media as well. Like Kotecha, Shaw framed death of Huntley through the truism that the latter was, after all, ‘always going to be a target because of the crimes he committed’.

I wonder how many viewers of the BBC news on 7 March watched this item to the end and thus heard the discussion at the end of what happened to Huntley at HMP Frankland, and how many, like me, caught or noticed only the first part in which nothing at all was said about it. The programme editors may rightly believe that making Huntley’s crimes in Soham in 2002 the first and central focus was a necessary reminder of why his death is newsworthy now, and to ensure due respect is paid to the memory of the child victims, their families, and members of the community affected.

I believe the BBC’s presentation of the item, and especially the text of initial report, throws some troubling light on our ‘cultural imaginary’ of crime and punishment.

The first is the way it speaks to a ‘displacement’ of justice into informal channels which, while often frighteningly uncontrolled and brutally violent, may not be entirely unpredictable or even necessarily intolerable. The presentation of Huntley’s case, for example, conveys an implicit expectation that   vengeful emotions will from time to time escape from their official (legal) constraints, and that this is itself part of what is popularly understood as ‘justice’ for society’s most reviled characters.

The release of Huntley’s girlfriend Maxine Carr with a new and secret identity, for example, precipitated several incidents of other women being mistaken for her and being relentlessly harassed, threatened, and abused by people who believed that ‘justice’ in her case too required much more than a few years in prison.

As for Huntley: ‘He was slashed across the throat in 2010, needing 21 stiches. The man who senselessly stole the futures of two happy, carefree girls…’ The juxtaposition of those sentences may not quite amount to tacit approval of the murder of a child-murderer. But they do seem to affirm a natural causal relation, and insinuate that the two crimes so separated in time represent a rebalancing of the order of things in a way not antithetical to justice’s values of desert and retribution. After all, the ‘retributive’ justification for punishment is retrospective, and the report – in its determination to focus only on that ‘ordinary Sunday in August’ of 2002 – implicitly participates in that punitive orientation back to the original crime.

The second point, which flows from this retrospective quality of the report, is the theme of genre and how genre tropes and conventions are used in ways that tell us something about the imagined role and responsibility of law and criminal justice. When I reflect on my own surprise at the lack of focus on Huntley’s death in the initial BBC report, I realise that this was the result of certain generic assumptions that I had made – mistakenly, as it turned out – upon hearing the initial headline.

I had imagined that the piece would follow the genre conventions of crime reporting, hence my expectation to hear details about the crime that had just been discovered taking precedence over the crimes committed 24 years previously by its victim. Instead, the report invokes a trope from an entirely different genre and thus a wholly different set of conventions, namely the retrospective: a staple of genres such as the memoir, biography (or in this case ‘geography’), and obituary. There may be various reasons for this, but I suggest that it allows the report to reflect a view about the ‘proper’ role of law with respect to ‘doing justice’ and its limits in respect of the ‘underserving’.

Notice that the vast bulk of detail given in the report concerns law’s positive and energetic involvement action in service of the ‘real’ victims of the ‘happy, carefree girls’, their families, and the local community of Soham, some of which Shaw then repeated in his interview:

‘The search lasted 13 days … it became one of the biggest operations of its kind … Huntley was arrested the same day … sentenced to life with a minimum of 40 years … Maxine Carr, who’d given him a false alibi, was sentenced to three and a half years …’.

By comparison, law’s involvement in Huntley’s death is not only minimised, but almost entirely set apart from it.

He was ‘always (going to be) a target’ on account of his own ‘senseless’ crimes, and an object of such understandable and widely shared hatred that his own violent end at the hand of a determined executioner could hardly be delayed forever.

The US Supreme Court Decision on Tariffs: Bad News from a WTO Perspective

Michail Risvas, 30 March 2026    8 mins read

On 20 February 2026, the US Supreme Court held in Learning Resources Inc et al v Trump et al that the International Emergency Economic Powers Act 1977 (IEEPA) does not authorise the President to impose tariffs, as he had done on ‘Liberation Day’. Although the decision of the Supreme Court appeared to provide a check on the President’s powers, its practical consequences were bound to be limited. However, the judges’ reasoning raises much wider questions about their skeptical perspective on international trade law.

As correctly recognised by commentators as well as Justice Kavanaugh’s dissent, ‘the decision might not substantially constrain a President’s ability to order tariffs going forward’, given  that numerous other federal statutes authorize the President to impose tariffs and might justify most (if not all) of the tariffs at issue in this case.’ Indeed, immediately following the ruling, the Trump Administration imposed a 10 per cent tariff under the Trade Act 1974 (s122) and initiated investigations into unfair trade practices by foreign countries under Section 301, thereby underscoring the limited practical impact of excluding IEEPA as a legal basis. Given the limited practical effect of the Supreme Court judgment on international trade, it is worth examining the symbolic significance of the decision for international trade law.

The IEEPA gives the President broad authority to regulate and block a wide range of international transactions in response to national emergencies:

‘investigate, block during the pendency of an investigation, regulate, direct and compel, nullify, void, prevent or prohibit, any acquisition, holding, withholding, use, transfer, withdrawal, transportation, importation or exportation of, or dealing in, or exercising any right, power, or privilege with respect to, or transactions involving, any property in which any foreign country or a national thereof has any interest by any person, or with respect to any property, subject to the jurisdiction of the United States.’ (emphasis added)

All six judges of the majority rejected the President’s reading of the IEEPA under ordinary principles of statutory construction that IEEPA’s grant of authority to ‘regulate… importation’ of property during a declared national emergency is broad enough to include the power to impose tariffs (that is duties on imports). Three of them, Justices Kagan, Sotomayor, and Jackson, – all appointed by democrat presidents – also relied on the ‘Major Questions Doctrine’ to bolster their conclusion. Under the doctrine, Congress must speak clearly if it wishes to assign to the Executive Branch decisions of vast economic and political significance.

The three dissenting judges, Justices Thomas, Alito, and Kavanaugh, who are amongst the most conservative members of the Supreme Court, held that the IEEPA included and delegated the power to impose tariffs.

The focus here will be on the disagreement between the majority and minority on whether IEEPA’s regulation of imports (or ‘importation’ in the Act) includes the power to impose tariffs from an international economic law perspective. However, the present analysis is not concerned with the issue of whether the Supreme Court was in fact required to consider international law. The conditions under which US courts engage with international law are limited (note, for example the Charming Betsy canon) and reflect a jurisprudence that ties the internalisation of international law to Congressional intent – in contrast the more expansive approach of the Court of Justice of the European Union (De Búrca (2015) at p. 1005).

An International Economic Law perspective on trade and tariffs in the Supreme Court decision

As a matter of international trade law, there are two important points in the majority’s reasoning. First, for the majority the concept of ‘regulation’ does not include tariffs, and this was not only a matter of statutory interpretation and Congressional practice but also a matter of common understanding. For example, Justice Kagan held that ‘the meaning of “regulate,” both in common parlance and as Congress uses the word, does not encompass taxing.’ By contrast, Justice Kavanaugh in his dissent emphasised ‘the dictionary definitions and the traditional understanding of tariffs as a tool to regulate foreign imports.’

Second and interrelated, Justice Roberts (who wrote the opinion of the Court) interprets tariffs as ‘very clear[ly]… a branch of the taxing power,’ ‘different in kind, not degree’ from regulatory tools, and thus outside the spectrum of authorities to ‘regulate’ trade. Again, Justice Kavanaugh disagreed characterising tariffs as a lesser form of trade regulation. He explains that:

‘[t]he plaintiffs and the Court acknowledge that IEEPA authorizes the President to impose quotas or embargoes on foreign imports—meaning that a President could completely block some or all imports. But they say that IEEPA does not authorize the President to employ the lesser power of tariffs, which simply condition imports on a payment. As they interpret the statute, the President could, for example, block all imports from China but cannot order even a $1 tariff on goods imported from China.’

These disagreements reflect deeper divergences on domestic constitutional law and what the imperative of the separation of powers requires. Notably, it is the majority’s reasoning that has been subject to criticism under doctrinal domestic law. For instance, Chad Squitieri has argued that ‘[t]o ignore that the power to regulate importation has traditionally included the power to impose tariffs, and to instead shift attention to whether the statute delegates a separate taxation power, is to miss the point.’ That point is that tariffs can and do regulate, quite regardless of whether they also raise revenue.

For present purposes, the focus is on how the Supreme Court majority’s understanding of tariffs sits uneasily with established WTO law and practice and, more broadly, international economic law doctrine. First, from an international economic law perspective, tariffs fall within the scope of trade ‘regulation.’ For example, Article XXIV of the General Agreement on Tariffs and Trade (GATT) (both in the 1947 and the 1994 versions) refers to ‘tariffs or other regulations of commerce’ and ‘duties or other regulations of commerce’ (emphasis added). This language makes clear that tariffs are not conceived as external to ‘regulation’ but rather as one of its central expressions.

Second, in the context of the WTO, tariffs not only constitute the paradigmatic but also the preferred regulatory tool in matters of international trade, as opposed to more restrictive regulatory measures such as quantitative restrictions which are prohibited. A WTO Panel in Turkey – Textiles  emphasized that ‘[t]he prohibition against quantitative restrictions is a reflection that tariffs are GATT’s border protection “of choice”. Quantitative restrictions impose absolute limits on imports, while tariffs do not’ [9.63].

Quoting this the Appellate Body in India – Additional Import Duties also held that ‘tariffs are legitimate instruments to accomplish certain trade policy or other objectives such as to generate fiscal revenue. Indeed, under the GATT 1994, they are the preferred trade policy instrument, whereas quantitative restrictions are in principle prohibited’ [159] (emphasis added).

Accordingly, the majority’s sharp conceptual separation between taxation and regulation does not align with the functional and doctrinal approach adopted in WTO law, where tariffs operate simultaneously as fiscal instruments and as central tools of trade regulation.

This raises the question why the progressive justices (in addition to three conservative judges), rather than the ‘conservative’ dissent, arrived at a position that is so antithetical to established WTO law and practice. One possible explanation is that the majority’s primary concern was to impose limits on the President’s powers as a matter of domestic constitutional law, even if this came at the expense of alignment with international economic law. Yet, the longstanding US opposition to the WTO Appellate Body – across the administrations of Presidents Obama, Trump, and Biden, and ultimately leading to its paralysis – also provides useful background explanation.

The judgment’s silence on contemporary international trade law and practice

In a decision that spans approximately 170 pages and examines the historical evolution of early modern English law and 18th and 19th century US practice, one might reasonably have expected some engagement with contemporary international trade law and practice. Even a brief acknowledgment of how tariffs are conceptualised in international trade law would have provided a more contextually grounded account of the regulatory landscape within which such measures operate.

The merits of the Supreme Court’s decision will remain contested from both constitutional and policy perspectives. However, irrespective of whether the Court was formally required to engage with international or WTO law, its complete silence on these frameworks is likely to reinforce the perception that scepticism toward international law and multilateral institutions, such as the WTO, is – within the United States – systemic rather than contingent, and not confined to any particular administration or ideological orientation.

Are there no ‘deserving poor’ anymore?

Lisa Whitehouse, 1 April 2026    7 mins read

It is nearly sixty years since Cathy Come Home first aired on the BBC. Written by Jeremy Sandford and directed by Ken Loach, this hard-hitting, documentary-style portrayal of a couple who, through no fault of their own, spiralled into homelessness and destitution, sparked national debate about the then housing crisis. The message arising out of the play was clear and unequivocal. The protagonists, Cathy and Reg were, in a manner reminiscent of the Victorian trope of the ‘deserving and undeserving poor’, worthy of our sympathy.

The changed neoliberal narrative: no rights without responsibilities

A lot has changed over the last sixty years, to the extent that we might now question whether Cathy Come Home would provoke the same response it did back in November 1966. The rise of neoliberalism post-1979, Brexit, and the growing popularity of right-wing political parties (among so many other things), have seen policy, rhetoric, and the media (print and social), cast certain categories of individual, such as single mothers, immigrants, and welfare recipients, as different, deviant and dangerous in the eyes of hard-working ‘responsible’ taxpayers.

Some have argued that this is a tactic used by proponents of neoliberalism to maintain support for its aims, categorising the welfare state as a cause of, rather than a solution to, poverty. The use of exclusionary and ‘othering’ language to depict those who are ‘them’ and not ‘us’ deflects attention away from the failings of the state and instead recasts it as a failure of individual responsibility. This responsibilisation narrative arises out of the reconfiguring of the relationship between the state and its citizens, particularly in respect of the provision of civil entitlements.

The creation of the welfare state in the UK, initiated by the Beveridge Report of 1942, was intended to create a system that supported citizens from the ‘cradle to the grave’. Some governments have, however, since the 1970s, sought to transform the state’s role from the guarantor of a universal entitlement to social rights, to a paternalistic and supervisory one that demands that citizens earn civil entitlements by conforming to expected standards of behaviour.

This new narrative insists that there should be ‘no rights without responsibilities’, thereby bringing an end to the perceived passive receipt of welfare benefits and justifying the exclusion of some from the social safety net if they fail to act ‘responsibly’. This reconceptualization of citizenship has been criticized on several grounds including its failure to recognise the impact that structural disadvantage and inequality, such as the shortage of affordable housing that led to Cathy and Reg’s downfall, have on the ability of individuals to act responsibly.

In this new era of individualised responsibility, it is possible that Cathy and Reg might be portrayed as undeserving, feckless, architects of their own downfall, and a drain on the responsible taxpayer. This narrative is not new. Smiles’ best-selling book of 1859, for example, promoted the moralistic view that debt, destitution and dependence were due to the failure on the part of the individual to exhibit the Victorian virtues of hard work, perseverance, and temperance. Smiles advocated ‘self-help’ as the remedy for poverty, encouraging, particularly young working-class men, to elevate themselves out of poverty through hard work, thereby ensuring their own happiness and wellbeing.

And yet, of course, we know that poverty is not a choice. Rather, it is influenced by individual and structural factors, many of which are beyond the control of the individual. Insecure and low-paid employment, high housing costs, individual life events, and an ineffective and sometimes punitive welfare system are just some of the causes of poverty.

The state’s complicity in facilitating inequality

It might be assumed that the state is best placed to tackle some of these structural constraints. However, what becomes apparent from a review of some public policy initiatives post-1979, is that rather than attempting to address structural inequality, the state has been complicit in facilitating it. A cynic might even suggest that the state has engaged in performative and populist policy-making that sets some individuals up to fail.

Take the Removal of the Spare Room Subsidy (RSRS), otherwise known as the ‘bedroom tax’. This reduces the rent that can be covered by welfare payments if social tenants reside in properties considered larger than they need. As of November 2025, 12,000 Housing Benefit claimants had a reduction to their weekly award amount due to the RSRS scheme. Up to date figures on the proportion of households in receipt of the housing element of Universal Credit who received a reduction due to the RSRS are not available, but to give some idea of the scale, in November 2020 it was 240,000.

The RSRS was intended to achieve behavioural change, encouraging or forcing households to move to smaller dwellings. However, this assumes that social tenants have a meaningful choice over where they live when, in reality, social housing has been decimated by policies such as the ‘right to buy’. Described as one of the ‘largest giveaways in UK history’, the policy allowed social tenants of sufficient standing to buy their council-owned home at a significant discount. Since its introduction in the early 1980s, around two million council properties have been sold, contributing to, what is now, a chronic shortage in and long waiting lists for social housing.

Importantly, however, the state was aware at the time it introduced the RSRS that there were insufficient properties of a smaller size to accommodate households subject to the deduction. For some of these households, this will have increased their dependency on the state, with evidence suggesting that many sought discretionary housing payments to cover the shortfall in their welfare payments. It would seem, therefore, that the state’s attempt to encourage welfare recipients to move to smaller accommodation has instead led to the state subsidizing those households to remain in their current home.

The unintended consequences of well-intentioned policies?

A less cynical view might suggest that inequality outcomes are the product of ambiguity neglect, a failure by policymakers to appreciate the potential for unintended consequences arising out of policies. These outcomes can often be generated by perverse incentives whereby citizens are incentivised to act in a manner that contradicts the intended outcome of the policy.

It is argued here, however, that in relation to aspects of policymaking since the late 1970s, a more apt label might be ‘perversity neglect’, constituting a failure by policymakers to appreciate the potential for perverse outcomes arising out of policies. In turn, rather than being incentivised to disrupt policy outcomes, individuals are subject to ‘perverse disincentives’ that hinder or prevent genuine attempts to comply with the intended outcomes of policies.

To take one example from the housing context, a policy intended to encourage social tenants to take responsibility for paying their own rent (by removing the direct payment of the housing element of welfare payments to landlords) led to some being unable to pay their rent (and therefore at threat of eviction) because of the mandatory five week waiting period before receipt of the first welfare payment.

Reckless policymaking driven by populist political agendas

The question arises as to whether the state’s complicity in generating structural inequality is intentional. That would, of course, be difficult to prove, but it can be argued that, if not intentional, then the state it is at the very least neglectful, if not reckless, in failing to appreciate the potential for perverse outcomes or, once aware of such, failing to address them. The answer offered up by research into ambiguity neglect is that the state should conduct evidence-based decision making prior to implementing policies, and post-implementation, to assess the outcomes of such, to identify and if necessary address unintended or perverse outcomes.

Some might argue, however, that neoliberalism and governments that engage in populist policymaking are not concerned with the outcomes of policies, or the likely success of the measures they give rise to, but rather with the perception that those policies engender (e.g. ‘stop the boats’). The perception that the welfare state rewards idleness, for example, has been used to justify the imposition of conditions on welfare claimants in order to make welfare ‘like work’. The outcome of the conditional nature of welfare can, however, in some cases, perpetuate poverty and dependency, including discouraging some claimants from entering into work in order to avoid having their welfare benefits reduced.

It is therefore questionable whether there is any motivation or sufficient public support for (particularly neoliberal) governments to avoid or remedy perversity neglect, particularly when there appear to be no ‘deserving poor’.

Palestine Action and the Paradoxical Status of Protests in Democracies

Alun Gibbs, 19 March 2026    8 mins read

Protests are never very comfortable. They may be disruptive, awkward and costly. In a democracy they also pose difficult foundational questions as they imply that change (sometimes quite radical change) is required and that this can be achieved only by a recourse to measures and methods which lie outside the usual democratic political processes, often for a variety of complex reasons. At the same time for a democratic state the fact of protest is not necessarily an indication of the failure of the political but can be a sign of the vitality and health of the constitutional life of a state – free expression and assembly are core rights which are invoked and ought to be upheld by protest.

In legitimate protest we glimpse the foundational engagement of our political association – popular causes – that foster a sense of common ground; so many constitutional states often forge their first breath within the ferment of protest and in the UK we can come face to face with the historical contingency of our freedom (Chartism, suffrage, employment rights and equality) by reflecting on the history of protest and its milestones. Protest then is paradoxical in character – it can emerge out of the feelings of rage at the political democratic process for its failures to address a burning concern, but it is also the sign of life, health and perhaps progress within a democratic political community.

The paradox facing the state around protest finds its way into the legal architecture as the fair balance between the state as the ‘policer’ of law and order and the state as ‘guarantor’ of the civil rights of freedom of association and assembly. How do we expect the balance to be struck was one of the key issues and concerns surrounding the decision of the High Court (13 Feb 2026) in the high-profile Palestine Action case.

In the early hours of 20 June 2025 several activists of Palestine Action, a direct-action protest group founded in 2020, broke into the RAF base at Brize Norten and spray painted two aircraft causing serious damage. In response, the Home Secretary decided to proscribe the organisation under the Terrorism Act 2000 as ‘concerned in terrorism’.

A terrorist organisation is defined as one that engages in, prepares, promotes or encourages terrorism, which in turn requires (1) a serious violation against a person, serious damage to property (as in this case), or similar, or the threat thereof; (2) which is designed to influence the government or intimidate the public; (3) for the purpose of advancing a political, religious, racial or ideological cause. Most of the 84 organisations so far proscribed under the Act are extremist organisations engaged in high levels of violence, typically Islamist extremists and white supremacists, and indisputably terrorist. There was significant outcry that the activities of Palestine Action should be aligned to those of such terrorist organisations.

The government decision seemed to be a reaction to the Brize Norten incident, but Palestine Action had been on its radar since 2024. There had been an escalating series of incidents targeting arms manufacturers in the UK linked to Isreal, such as Thales, a French space and defence corporation, and Elbit Systems, Israel’s largest armament company, each time involving serious damage to property.  

As consequence of the group’s proscription (ban), it also became a criminal offence to offer or express opinion in support of it. In the months that followed more than 2000 supporters had been arrested for holding banners in support of Palestine Action, attracting the attention of media both in this country and around the world and prompting concern as to the health of the right to protest in the UK.

Huda Ammori, one of the founders of Palestine Action, brought a challenge against the government’s decision to ban the organisation. The High Court decision constitutes the opening salvo in what is likely to attract further litigation. It decided that the Home Secretary’s determination was unlawful but declined to quash the order – meaning that Palestine Action remains banned (presumably pending the outcome of further legal action). It leaves things in legal limbo, especially for those that have already been arrested.

Taking the activities of Palestine Action as a whole and including the most violent and high-profile instances met the threshold of the statutory test for proscription. The court also noted that Palestine Action was not engaged in civil disobedience, as set out in Cuadrilla Bowland Limited v Persons Unknown (2020), given that the activities of the group were unrestrained, dangerous and criminal in nature. So, the outcome of this case is not a vindication of Palestine Action or its methods.

Even though the statutory terrorism test was met, it does not necessarily mean that Palestine Action should have been proscribed. The ‘policy’, devised by the Secretary of State, is to determine whether it is proportionate and necessary to proscribe the organisation. As the High Court explained, the core of the policy for exercising her discretion is that ‘not all organisations that meet the concerned terrorism requirement should be proscribed.’

The court found that the reason behind the proscription was to subject Palestine Action to the criminal consequences of the proscription which would be severe enough to disrupt or prevent the group from continuing. As these consequences would always and inevitably be the effect of proscription this could not, in the court’s view, be considered a relevant factor within the policy.

If this seems like rather a narrow basis to hold that the determination was unlawful, the court was somewhat more expansive in the consideration of the proportionality of proscription in relation to Article 10 and 11 rights in the Human Rights Act 1998, freedom of association and freedom of assembly. Once proscription has taken place there is obviously a considerable interference with Articles 10 and 11 rights given that it prevents any further protest, activity or support.

On this point the court considered that the activities of Palestine Action had ‘not yet reached the level, scale and persistence that would justify the application of the criminal law measures that are the consequence of proscription’ – even though the court acknowledged that Palestine Action had been involved in unrestrained and disproportionate criminal activity, and not civil disobedience. In other words, even if the Home Secretary had followed her policy on proscription, it would still fail the proportionality test when it came to the rights protected by the HRA.

Usually on national security issues, such as this, the court would adopt a more deferential attitude to the view of the executive on where the balance lies. Thus, it is important that – in terms of the wider architecture of protest and civil liberties – the court adopted an independent view regarding the fair balance, given that proscription can (as in this case) have a significant impact on the enjoyment of convention rights. How far this analysis of the balance will be upheld in subsequent litigation remains to be seen.

In this wider sense the ruling can be viewed as taking a tentative step to protect this understanding of the place of protest by recognising that the law ought to adopt a sceptical attitude to powers of proscription. It set a reasonably questioning approach for the executive to show that a fair balance is in fact attained when it comes to the limitations of rights of freedom of assembly and association. It is only right that the courts should be willing to check the powers of the executive in this way so that anti-terrorism legislation cannot be used to unlawfully suppress the conduct of political debate and the legitimate rights associated with it.

At the heart of this debate about civil disobedience and protest we see the contours of a much more deep-seated issue of constitutional life. The authority of the state is not something that exists independently of the individual’s moral perspective. Rather, it is also grounded in the willingness of the state to protect the capacity and exercise of personal moral judgment. Any claim that the State has the right to command, and that citizens have a duty to obey, ultimately rests on creating a space for moral judgements in relation to the law as part and parcel of the wider claim of the law that it rests on reason rather than simply an act of will. This principle implies that the moral autonomy of judgement is not external to law but internal to it. The very idea that a law can bind us presupposes that we are the kind of agents who can understand, interpret, and morally assess reasons for compliance. A citizen is not merely a passive recipient of commands but an active participant in determining whether those commands carry legitimate authority and is not simply an act of coercion. The place of protest and its legal protection is one such constitutional space that serves this purpose within constitutional life and as such infringements on it must be viewed critically and limitations on it examined cautiously. It is within this wider spirit that we must look at the current case regarding the proscription of Palestine Action.

Trump’s takeover of Venezuela and the long-standing hypocrisy of international law

Andrea Maria Pelliconi, 5 February 2026    7 mins read

The past months have brought back with startling clarity a pattern many international lawyers know all too well. Under the second presidency of Donald Trump, the United States (US) has returned to overt coercion to impose their interests upon the rest of the world. Realist pragmatism has always been present in international relations, but states used to couple their acts “with at least a resemblance of legal justification”. This time, however, practice is matched by words: Trump has completely dispensed with the liberal varnish that usually accompanies US extraterritorial mischiefs and openly admits that his actions are guided only by his own morality. The attack on Venezuela and the kidnap of Nicolás Maduro and his wife have been followed by repeated threats to annex Greenland by purchase or use of force, alongside renewed intimidation directed at other states such as Panama, Mexico, Colombia, and Cuba. We’re witnessing an accelerated deepening of the global crisis of multilateralism and international law and a return to Great Powers’ “spheres of influence”.

Everyone knows in legal scholarship that the unilateral kidnap of a sitting head of state is unequivocally unlawful under international law, whatever one may think about the Maduro regime and its gross and widespread human rights abuses. The prohibition of the use of force, the principles of sovereignty and non-intervention, and the personal immunity of heads of state vis-à-vis other domestic jurisdictions leave no room for ambiguity. And yet, the responses of some western states and the European Union (EU) have not been so unequivocal. Instead, they have been divided and ambivalent, cloaked in watered-down words of “concern” or “monitoring” or strategic silences. German Chancellor Merz has stated that the legal assessment of the US operation is “complex”. French President Macron has emphasised the need for a transition which is “peaceful, democratic, and respectful of the will of the Venezuelan people”. Italian Prime Minister Giorgia Meloni suggested that while the use of force is generally wrong, the US were acting in self-defence against so-called “hybrid security attacks”, referring to Maduro’s supposed weaponisation of drug-trafficking against the US. Reactions were far more decisive when it came to Trump’s threats against Greenland, where the territorial integrity and interests of an EU and NATO state were at stake. European leaders suddenly rediscovered the language of international law and the inviolability of sovereignty and territorial integrity with admirable clarity.

This should not come as a surprise. Over the past years, manifestations of double standards and selectivity have grown exponentially, with the most discussed example being the reactions to Russia’s invasion of Ukraine and Israel’s “plausible” genocide in Gaza, both framed as defensive responses to security threats and terrorism. While one may well argue that Israel’s self-defence claim was more well-founded than Russia’s, it soon became clear that the 7 October 2023 attack was being used as a mere pretext to unleash uncontrolled violence on Palestinians, while preparing the ground for territorial expansion in both Gaza and the West Bank. This expansionist plan continues to develop even now, after the supposed “ceasefire” and the UN Resolution on Gaza, and nothing concrete is being done to bring it to an end. The same double standard surfaced in reactions to the International Criminal Court’s (ICC) arrest warrants against Vladimir Putin and Benjamin Netanyahu for international crimes. The same European states that rushed to praise the ICC for its investigation of the Russian leader, later declared that Netanyahu enjoys immunity from ICC prosecution.

And yet, once again, this selectivity is not new at all. It follows decades of US (and broader Western) unlawful attacks and interventions in Iraq, Afghanistan, Iran – to the extent that exceptionalism is not exceptional anymore. Each time, international law norms were stretched, re-interpreted, or reinvented through securitisation doctrines such as preventive self-defence and the “unable or unwilling” test, or hidden behind ostensible “benevolent motives” such as humanitarian intervention, regime change, and the “exportation of democracy”. In fact, the US has even committed the exact same type of head-of-state kidnapping before in the region, most notably with the capture of Manuel Noriega from Panama. This is the so called “rules-based international order”, meaning the rules that the US and its allies imposed upon the rest of the world (“the West and the rest”), as opposed to what international law actually required. Every time, other western states have been weak in condemning the illegality of these actions, and even weaker in doing anything concrete to prevent, stop, or redress them.

Now, at what feels like the climax of the collapse of the international legal order and multilateralism, everyone has awakened, shouting that this is not a drill. UN experts warn that such actions normalise lawlessness in international relations, and commentators caution that Venezuela sets a dangerous precedent: if powerful states may unilaterally decide when international law applies and when it does not, the legal order collapses into selective enforcement and strategic convenience. Even leading US academics now talk about the catastrophic collapse of jus ad bellum norms and the dangers we all face when “might unmakes right”. They highlight the risk that the Venezuela incident “opens the door to other similar actions by powerful nations in the future”. This fails to appreciate that the door has long been wide open.

A prevalent position now is to acknowledge the flaws of international law but vehemently opposing the abandoning of its normative constraints because they’re the only thing that will save us from debacle. International law can still be mobilised to place constraints on power and if it’s consistently disregarded it’s because of contingent political factors militating against full compliance. A recurring metaphor that I’ve heard a few times lately, including among critical thinkers, is the Sisyphus’ myth: the futility of the task should not deter the discipline. But this, at least in part, obscures how this system was intentionally built to shield the actions of the West and has laid the foundations for the situation we now find ourselves in. International law and the indeterminacy of its content provide a “professional vocabulary” to build plausible arguments. For decades, mainstream scholars were complicit in the legal legitimisation of these actions, coming up with doctrines that served the interests of the moment under a façade of international legal jargon.

The inherent defect of the international legal infrastructure has simply become more visible now. As Rajagopal has put it, “The revival of overt colonial and imperial designs under the Trump regime in Washington is notable not because it has invented new forms of domination, but because it has dispensed with the traditional liberal rhetoric that once accompanied them.” From its colonial origins to its modern doctrines of sovereignty, intervention, and trade, international law has consistently operated in the interests of dominant states and classes, while insulating them from the equal application of its norms. What we are witnessing today are colonial revivals as the logical outcome of a system that never truly decolonised. It seems scarier to European eyes now because it has eventually turned against them.

Of course, legal scholarship is not monolithic. Critical voices, including TWAIL scholars, have raised these concerns all along. Yet they have been unable to bring material change, partly because of the structural hierarchies of international law, and partly because of fragmentation of their own views. Some tried to change the system from within; others were content to critique from the margins; others advocated radical transformation or the complete dismantling of the legal order, often without a clear project – and always with different opinions – for what should come after.

Now, as we stand on the verge of a concrete dismantling of the system, with Trump’s plan to replace the United Nations with his own personal “Board of Peace” and international relations reaching the peak of personalisation and corporatisation, everyone – even critics – seem unsure what to do. If there is a moment to seize, it is now. But seizing it requires more than lamenting Trump’s excesses or the fragility of the system: it demands an honest reckoning with the errors of the past, Western exceptionalism, legal complicity, resource-hungry capitalism, and a system ostensibly built on sovereign equality but consistently seized by vetoes and unilateral reprisals. What is needed are visionary ideas for radical change and possible futures – and I am not sure we – myself in primis – are up for the task.

The Sense and Nonsense of AI Ethics: a Whistle Stop Tour

Kieron O’Hara, 29 January 2026 – 8 mins read

Emeritus fellow, University of Southampton, kmoh@soton.ac.uk

AI research has been energised since the unveiling of AlphaGo in 2016 and ChatGPT in 2022, demonstrating capabilities well beyond public and even expert expectations. It also has acquired a chaperone, a growing cottage industry of AI ethics to describe, diagnose, and ultimately remedy its perceived potential harms.

The Silicon Valley credo ‘move fast and break things’ is obviously ethically flawed (especially when the ‘things’ are people), but potential problems don’t usually spawn sub-disciplines; there are no ethics of differential equations, printing or hammers. There is no legal demand to be ethical, and no-one can force you to be ethical, so there is a limit to the number of harms it can prevent. That is not to say that AI development has no ethical dimension; of course it does, and I shall sketch it at the end of this blog.

The cottage industry has emerged from two sets of incentives. Ethicists like advisory committees (upon which they might expect to sit). A lovely example is Henry Kissinger, Eric Schmidt and Daniel Huttenlocher, who prescribe “the leadership of a small group of respected figures from the highest levels of government, business, and academia” to ensure the US “remains intellectually and strategically competitive in AI” and “raise awareness of the cultural implications AI produces”. I wonder who the former Secretary of State, the former CEO of Google and the MIT computer scientist have in mind? Meanwhile, tech developers relish applying Silicon Valley methods to moral philosophy, preferring doomster sci-fi to the hard yards of solving genuine problems (and if ethical codes raise compliance costs for startup competitors, what’s not to like?).

The result is a crowded field with a confusion of non-problems, non-serious problems, non-specific problems and the real deal. Apologies for my necessarily cursory treatment in this survey.

Non-problems

Some perceived AI ethics problems require little action beyond an eyeroll. One non-problem is Artificial General Intelligence, the singularity, and sentience, which together supposedly pose an existential threat. It is assumed without proof that superintelligent agents will have the power (and inclination) to pursue harmful goals autonomously. Barring thought experiments, game theory and the plot of 2001, no evidence is produced for this, although one expert declared, with spurious precision, that AI will take over in December 2027. Both Hinton and Bengio claim the risk is 20% or more. Is it really more serious than climate change, another pandemic or nuclear war?

A second type of non-problem uses critical theory to depict AI as complicit in capitalism, whiteness, racism, sexism, data colonialism, and so on. Maybe, maybe not, but it is not obvious what the conscientious AI developer is to do, other than indulge in the Foucault-worshipping armchair revolutionary groupthink that has thus far proved remarkably unsuccessful in derailing capitalism.

Non-serious problems

Some genuine problems may be poorly framed, not necessarily trivial or easily solved, but not an ethical priority either.

One such is bias. Because algorithms uncover existing patterns in data, they need either unbiased or synthetic data. If unbiased data is unavailable or insufficient, anticipate potential problems and biasing the algorithm against unwanted results. This may not be easy, but it’s a practical problem. Undesigned biases of algorithms are of less ethical import, being statistically anomalous rather than socially significant. Bias is additionally misleadingly framed by the disparate impact doctrine. This ignores that unintentionally discriminating decision procedures often discriminate for desired behaviour, such as being creditworthy or law-abiding. AI’s potential depends on positive discrimination; focusing only on negative discrimination is itself biased.

A second problem is that neural nets are black boxes; difficult if we demand that the machine explains its own output. But any AI ‘decision’ is implemented by an organisation, with responsibility to justify its actions. The required explanation is less the derivation of the output, than its congruence with the organisation’s goals, and the legitimacy of pursuing them.

Third is the persistent yet so far unproven claim that AI will replace jobs, leading to a shortage of work. This accepts the lump of labour fallacy, denies that greater productivity will raise wages and employment, and assumes an elusive business model for generative AI. And just because a job could be taken does not mean it will be; train drivers could have been eliminated 50 years ago, but they still chug along.

Fourth, privacy: but if training uses personal data without data subjects’ consent or other data protection grounds, then it is illegal. If it uses personal data legally, then its being unethical is not a strong ground upon which to act. Either way, data protection law outweighs ethical considerations.

Non-specific problems

Problems not unique to AI are not best addressed through AI regulation or governance. A recent paper listed the likely harms of AI as biodiversity loss, carbon emissions, chemical waste, exploitation of gig workers, exploitation of marginalised and indigenous groups, widening inequalities, eroding trust, and injuries to animals. All serious, but AI-specific guidelines are neither necessary nor sufficient to deal with these far wider issues.

Misinformation is also a problem, but a war against fake news will be as problematic as the war on drugs for the same reason: the issue is not supply, but rather excess demand. This is a social problem, requiring societal adaptation.

The real problems

Real problems require ethical insight, and AI developers need some control. One such is information pollution. LLMs have a tendency to ‘hallucinate’, and can be ‘trained’ to produce racist or other offensive output. This is particularly problematic because output will be used to train the next generation of bots, with the danger of a vicious circle of ‘pollucination’. Conscientious developers, under pressure to produce ever more compelling models, may be urged to put power before rigour.

Other serious issues include intellectual property, cybersecurity, defence (using autonomous learning systems), and diversity (progressive young males tend to be overrepresented in development teams). With these, the question is how to make AI safer without compromising quality, e.g. by insisting on ‘humans in the loop’.

Approaches to avoid

Simplistic views translate complex ethical positions into simple calculi. Framed like this, AI can solve the problems itself! Examples include:

  • Accelerationism: AI is superior to human thought, and development should be escalated to eliminate ‘residual anthropolitical signature’.
  • Effective altruism: a combination of the naïve utilitarianism of Derek Parfit, Peter Singer, and William MacAskill, with the debatable assumption that the tech bros’ chatbots are to benefit humankind, not their bank balances.
  • Ethical AI: AI systems themselves compute the ethical consequences of proposed actions.
  • Rationalism: extreme technocratic and hyper-rational consequentialism, ignoring convention and taking ideas to their logical conclusions.

Others assume that as humanity is transformed, AI systems and cyborgs will have divergent interests from humans, and yet comparable ethical status.

  • Transhumanism: humanity should be improved by applying technology to cognition, well-being and longevity.
  • Posthumanism: humanity should be eliminated by applying technology to genetics and neural capabilities to integrate it with wider sentient networks.
  • Environmental ethics and other anti-anthropocentric views: humans should not be central to ethical inquiry.

We should reject these too: anthropocentricity is central to ethical inquiry. Even views taking technologies, other species of animal, or entire ecosystems into account do so for anthropocentric reasons.

The third class of simplistic views argues for inclusive democratic participation (often phrased in complicity with the reader, suggesting that ‘we’ should take charge). Quite how citizens’ juries and civic dialogues will avoid being dominated by the exam-passing classes, or could constrain AI development is left unsaid – and good luck if you want to try it in China or Russia.

Finally, AI is neither intrinsically good nor bad. Technology to support the development of innovative software, drugs or defences could equally produce new malware, poisons or weapons. This dilemma can’t be offset by programmes of ‘AI for good’. These, while demonstrating benefits (‘beneficial’ defined by developers), can’t eliminate harmful or criminal uses.

The literature on ensuring that the ‘values’ of AI systems align with those of wider society is similarly flawed. Autonomous AI systems operate on reinforcement functions, not values or internal motivations, which they don’t have. They may behave unpredictably, and against human interests, but that can’t be programmed out of them. Testing and modelling methods will be far more use.

Virtue and responsible AI

Where exactly should ethics feature in AI? Let us begin with an observation: if someone sincerely wants to avoid doing the wrong thing, and is thoughtful and conscientious about it, then more often than not they will succeed. The result of the inquiry is less important than its existence.

The ethical actor is not a corporation or an in-role manager (they need back-covering standardised tick-box templates), but individuals involved in the development of AI within a corporate or organisational context. What matters is their conscientiousness and sincerity – their character, their virtuousness.

Virtue ethics has the usual enforcement problem – no-one is forced to be virtuous, and sometimes virtuousness may be punished (cf. the Trump administration passim). However, it is normative for the ethically-minded, and credits the developer with trustworthiness and maturity.

There is a connection between virtue ethics and responsible AI, the movement to ensure ethical development through the life cycle of design, coding, training, and deployment. This literature, such as Vallor on virtues and Dignum on responsible AI, despite familiar Western and progressive biases, is the most valuable starting point for AI ethics.