The University of Southampton

Author: Ben Van Blerk

The Magnificent and Scary Future of Humanoid Robots

Posted on by
My idea of an automated home (AI-generated image)!

People have dreamed for decades about a world automated by robots so humans can just lie back and relax. Inspired by the idea of the ‘6 Million Dollar Man’, I found what I think may be the ultimate engineered replacement body part: an entire “human”. Humanoid robots mimic the human form, allowing them to carry out jobs in environments that are specifically built for humans. This is a growing market, predicted to increase to almost $38 billion by 2035, according to a Goldman Sachs report. With the rapid expansion of artificial intelligence (AI) as well, humanoid robots will likely have a glowing future.

A great example of a humanoid robot (and my favourite!) is Atlas, produced by Boston Dynamics. I’ve been a fan of Boston Dynamics’ robotics work for many years, so I had to give them a mention! Atlas was originally a hydraulic model designed for search and rescue operations in 2013, funded by the US Defense Advanced Research Projects Agency, but this model has since been retired and replaced by a fully electric, commercialised robot to be used in boring or more-importantly, dangerous work environments. Atlas is not yet available for purchase, but it shows great promise. This video from earlier this month showcases some of the new model’s dynamic abilities:

Combining humanoid robots with AI seems to be a pressing matter for this industry. AI models allow these robots to essentially train themselves, which helps them to develop and adapt to new situations more quickly. After starting my research into this blend, I watched the CEO of Nvidia, Jensen Huang, give a keynote speech in which he unveiled the world’s first open-source foundation model for the widespread use of humanoid robots, named “Isaac GR00T N1“, or ‘Groot N1’. This ‘generalist’ AI model is trained on both real and synthetic data, which makes humanoid robots more capable in a wider range of environments even with less training. With Groot N1 available to the public, along with blueprints and frameworks for synthetic training data, I believe this model will massively accelerate the production of capable humanoid robots.

Jensen Huang talks about the future of humanoid robots in March 2025 keynote speech.

I’ve always thought of humanoid robots as a force for good, but my research led me down a rabbit hole of ethics and fears surrounding their future. Here are just a few of the internet’s many concerns for you to think about:

  • Could humanoid robots act autonomously in ways that don’t align with human values?
  • If robots are capable of acting on their own accord, who is responsible for their actions?
  • Humanoid robots, like Atlas, rely on sensors and cameras in order to perceive their environments. Could this be an invasion of privacy?
  • Finally, if these robots develop to the extent of becoming ‘superintelligent’, could this lead to a future where the planet is controlled by machines?

In short, humanoid robots show great promise for creating my dream of an automated life by seamlessly fitting into environments created for humans. This is thanks in particular to the current efforts of Boston Dynamics with Atlas and Nvidia with Groot N1. However, I now realise that there are a whole host of important ethical and legal issues that should be discussed before the market for humanoid robots grows too large, to ensure our safety and privacy.

References:

Boston Dynamics. (2025). Atlas | Boston Dynamics. Available at: https://bostondynamics.com/atlas/ [Accessed 23 Mar. 2025].

Callari, T.C., Segate, R.V., Hubbard, E.-M., Daly, A. and Lohse, N. (2024). An ethical framework for human-robot collaboration for the future people-centric manufacturing: A collaborative endeavour with European subject-matter experts in ethics. Technology in Society, 78, p.102680. https://doi.org/10.1016/j.techsoc.2024.102680.

Du, J., Isayama, Y., Costa, D., Delaney, M., Zheng, N., Xu, O., Zhao, T., Li, Z., Chen, H. and Ye, Z. (2024). Humanoid Robot: The AI accelerant. Available at: https://www.goldmansachs.com/pdfs/insights/pages/gs-research/global-automation-humanoid-robot-the-ai-accelerant/report.pdf.

Huang, J. (2025). GTC March 2025 Keynote with NVIDIA CEO Jensen Huang. 18 March, SAP Center, California.

NVIDIA Developer. (2025). NVIDIA Isaac GR00T. Available at: https://developer.nvidia.com/isaac/gr00t [Accessed 23 Mar. 2025].

Obrenovic, B., Gu, X., Wang, G., Godinic, D. and Jakhongirov, I. (2024). Generative AI and human–robot interaction: implications and future agenda for business, society and ethics. AI & Society. https://doi.org/10.1007/s00146-024-01889-0.

Tong, Y., Liu, H. and Zhang, Z. (2024). Advancements in Humanoid Robots: A Comprehensive Review and Future Prospects. IEEE/CAA Journal of Automatica Sinica, 11(2), pp.301–328. https://doi.org/10.1109/jas.2023.124140.

‌

‌

Stem Cells and Parkinson’s: A New Hope for Treatment?

Posted on by

Parkinson’s disease is a neurogenerative disorder caused by the loss of nerve cells in a part of the brain called substantia nigra. Losing these cells causes a drop in the abundance of a vital chemical called dopamine, leading to a whole host of symptoms, including tremors, stiffness, and dementia. Parkinson’s disease can also strongly impact the mental health of affected people, often associated with depression and anxiety.

This disease is common, affecting more than 1 in 50 people over the age of 65, with some people even experiencing symptoms under 40 years old. The cause of the degeneration of specific nerve cells associated with Parkinson’s disease is unknown, but there is evidence for some new approaches for its treatment using stem cells.

How could stem cells be used?

First, what are stem cells? There are a few kinds of stem cell, but the term generally refers to cells that are capable of developing into other types of cell. Embryonic stem cells, in particular, are pluripotent, meaning they can give rise to all cell types. Induced pluripotent stem cells (iPSCs), which also differentiate into any cell type, can be created via the reprogramming of existing adult cells. Of all stem cells available, these two seem to be the most promising.

One major therapy which has the potential to treat Parkinson’s disease is “Bemdaneprocel”, or “BRT-DA01”. In this therapy, human embryonic stem cells are converted into dopamine-producing neuron progenitors in the lab before being surgically implanted into a specific area of the brain. These cells may be recognised as foreign, so patients must take immunosuppressants for a year. This method began a Phase 1 trial in 2021. “BRT-DA01” showed high safety and potential clinical benefits, and is expected to undergo a Phase 3 trial in the first half of this year.

Another experimental therapy for Parkinson’s disease which aims to replace lost dopaminergic neurons was approved for a Phase 1/2a trial in the US by the FDA in late 2023. In this method (“ANPD001”), skin cells are collected from a patient and reprogrammed in the lab, producing iPSCs which are then converted into immature dopamine neuronal precursor cells. Finally, they are transplanted into the brain, in the hope that they will develop into those crucial neurons. One key advantage of “ANPD001” over “BRT-DA01” is that iPSCs are derived from the patient, so immunosuppressants aren’t needed. As of January this year, this method has been safely carried out with two patient cohorts.

Some ethical issues

While embryonic stem cells may be an effective source for the treatment of Parkinson’s, testing and using them for science raises some interesting ethical questions. These cells are derived from early-stage embryos, in a process involving the destruction of the embryo. The moral status of the embryo in comparison to human life therefore must be questioned. Furthermore, embryos used for scientific purposes are often donated by people undergoing in vitro fertilisation. I feel it is crucial that donors are entirely aware of how their embryos will be used, and clear consent is given.

This is obviously a difficult ethical dilemma, though there seems to be a work-around. Fortunately, iPSCs avoid the ethical questions faced by embryonic stem cells, because they are not derived from embryos. As a result, these cells seem to be a far more acceptable choice for the treatment of Parkinson’s disease, especially from the standpoint of public concern.