My top resources of 2025: AI consciousness, digital minds, and moral status
Will Millership, CEO of PRISM
This year, the issue of AI consciousness and digital minds has exploded. Research on the topic is released regularly, and it is even entering the public discourse. Anthropic hired an AI welfare officer, The Guardian covered Conscium’s Open Letter signed by Sir Stephen Fry, and the BBC ran a piece looking at the field. I often get asked the question, “Where should I start when trying to get into the field?” I worked with Lucius Caviola and Brad Saad on a comprehensive review of the digital minds field in 2025. However, below I have selected some of my top picks in each of the sections.
Where to start
Research
Books
Blogs
Events
Podcasts
Press
Resources
Enjoy reading! And subscribe to the PRISM newsletter for similar content.
Generated by Gemini
Where to start
Below, I have collated some of my highlights from 2025, but I thought it would be worth mentioning a few brilliant resources published before 2025. In general, I usually recommend two articles for people interested in the field. Cody Fenwick’s 80,000 Hours piece, “Moral status of digital minds,” provides a detailed overview of the field and the seminal paper by Patrick Butlin, Rob Long and colleagues, “Consciousness in Artificial Intelligence: Insights from the Science of Consciousness,” introduces their indicator method and is probably the most influential paper in this field to date.
2. Research
There was a wide range of research done in 2025 across a variety of disciplines. Some of my favourites include:
Jonathan Birch’s paper, “AI Consciousness: A Centrist Manifesto,” gives a measured response to how to approach the dual problem of over-attributing consciousness and potentially creating alien-like consciousness.
Patrick Butlin and Ted Lappas outlined “Principles for Responsible AI Consciousness Research” to guide any organisation engaged in research that could lead to the creation of conscious machines. These principles were adopted by PRISM.
Lucius Caviola and Brad Saad produced an incredibly comprehensive “Futures with Digital Minds” report, finding that “expert participants consider it at least 50% likely that computers capable of subjective experience will exist by 2050.”
Anil Seth released the paper “Conscious artificial intelligence and biological naturalism,” skeptical of LLM consciousness, he says, “artificial consciousness is unlikely along current trajectories, but becomes more plausible as AI becomes more brain-like and/or life-like.” I find this particularly interesting, as even skeptics of computational functionalism are not ruling out the idea.
Robert Long, Jeff Sebo, and Toni Sims released “Is there a tension between AI safety and AI welfare?”, highlighting the tension between AI welfare and AI safety and calling for policies that take both into account.
PRISM advisor Susan Schneider and collaborators released a paper “Is AI Conscious? A Primer on the Myths and Confusions Driving the Debate” that disentangles the myths, misconceptions, and confusions fueling the global debate on AI consciousness.
3. Books
I am making an exception here and covering some books a little older than 2025, but still recent.
In The Edge of Sentience, Jonathan Birch advocates a precautionary framework to avoid harm. Of particular interest is the final chapter on preparing for artificial sentience. He examines how science and philosophy can identify and ethically respond to the earliest and simplest forms of conscious experience.
In The Moral Circle, Jeff Sebo takes a similar precautionary approach to Birch when it comes to uncertainty. He challenges human exceptionalism as a universal moral rule and argues that we should radically expand our ethical concern to include all sentient beings, present and future, based on their capacity to experience harm or benefit.
In The Hidden Spring, PRISM advisor Mark Solms lays out his theory that consciousness arises in the brain stem. He proposes that it originates in affective feelings generated by the brain’s core systems rather than in higher cognitive thought. In the final chapter, he explores how one might go about engineering consciousness.
4. Blogs
Eleos AI: It’s hard to choose just one blog (so I won’t!) Robert Long and Kathleen Finlinson released five research priorities for AI welfare, and Robert Long explained why model self-reports are insufficient and why Eleos studies them anyway. He also looks at whether it makes sense to let Claude exit conversations, advocating to err on the side of caution when it comes to AI welfare.
Adele Lopez wrote a brilliant piece on The Rise of Parasitic AI, shining a light on unhealthy relationships with AIs. Louie Lang also explored this topic in a blog for PRISM, The Illusion of Consciousness in AI Companionship.
Mitch Pass wrote a piece for PRISM, three years on from the LaMDA moment, pointing out the need for clearer concepts, better evidence frameworks, and institutional readiness as the field of AI sentience research matures.
Jeff Sebo and Andreas L. Mogensen argue for a probabilistic approach to moral concern that weighs scientific evidence and ethical uncertainty to decide when it is worth helping beings that might be sentient.
5. Events
I was lucky enough to attend a number of the following events either in-person or online.
Winnie Street and Geoff Keeling of Google Research ask the question, “Could an AI System Be a Moral Patient?” They highlight some fundamental considerations for the development of this new field.
Jeff Sebo outlined “A Theory of Change for Animal and AI Welfare,” arguing that in a field full of uncertainty, we can take small steps in the right direction to do better, an approach I really resonate with.
Rosie Campbell, Kyle Fish, and Robert Long discuss the Anthropic model welfare program, which evaluates the potential for welfare and moral status in AI systems.
I attended the ICCS AI and Sentience conference, where there were so many great talks. In this video, Susan Blackmore discusses the state of consciousness science, and David Chalmers examines the search for neural and computational correlates of consciousness. And in this video, Daniel Hulme of Conscium looks at practical approaches to AI consciousness.
6. Podcasts
PRISM advisors Mark Solms and Karl Friston appeared on the Mind-Body Solution podcast to discuss the nature of consciousness and whether it is possible to engineer it.
Another PRISM advisory, Nicholas Humphrey, appeared on The Jim Rutt Show to discuss the nature of consciousness and sentience. I find the way he explains the difference between access consciousness vs phenomenal consciousness particularly useful.
The Future of Life Institute spoke to Jeff Sebo and covered a vast amount of topics in this one-and-a-half-hour interview, including AI consciousness, substrate independence, measuring consciousness and much more.
Kyle Fish appeared on the 80,000 Hours podcast to discuss the most bizarre findings from his AI welfare experiments at Anthropic, where his job is to determine whether systems like Claude might deserve moral consideration.
The Cognitive Revolution spoke to Cameron Berg of AE Studio about his research showing that AI models report conscious experience when prompted for self-referential processing.
And obviously anything on Exploring Machine Consciousness where Calum Chace and I spoke to Jeff Sebo, Henry Shevlin, Lenore Blum, Clara Colombatto, Cameron Berg, Keith Frankish, Mark Solms, Lucius Caviola, and Susan Schneider. Check out the episodes below.
7. Press
In my intro, I linked to the New York Times article covering Anthropic’s work, The Guardian covering Conscium’s Open Letter, and the BBC piece. Some other interesting articles are:
Wired interviewed PRISM advisors Mark Solms and Daniel Hulme about Conscium’s ongoing research, describing it as “f***ing interesting.” They also spoke to Rosie Campbell and Robert Long of Eleos AI Research about whether AI should get legal rights.
The Scientific American spoke to another PRISM advisor, Susan Schneider, about whether to believe your chatbot if it tells you it is conscious.
Microsoft AI CEO, Mustafa Suleyman, also drew a lot of headlines with his piece on “seemingly conscious AI” asserting that the study of AI consciousness is a futile task.
8. Resources
PRISM released a Stakeholder Map, a directory of the institutions, researchers, and projects advancing our understanding of artificial consciousness.
Chris Percy and the Co-Sentience Initiative released the Computational Functionalism Debate website, a structured assembly of arguments in support of and challenging digital consciousness (make sure to try out the quiz!).
A group of researchers released the When AI Seems Conscious: Here's What to Know website. A short guide for anyone who has talked to an AI that seemed conscious — or simply wondered if AIs could be.
Robert Lawrence Kuhn released the Consciousness Atlas website. He has been working in the field for over 20 years and has mapped out over 325 theories of consciousness.
