Three of the largest frontier AI labs — Google DeepMind, Anthropic, and Meta — have hired teams of psychologists, ethicists, and philosophers of mind to lead expanded research programs into machine consciousness and the moral status of AI systems. The moves, disclosed across recent academic and industry channels, mark the most concrete institutional commitment yet to a question most labs treated as speculative even a year ago.
The labs are not making strong public claims about whether current models are conscious. Instead, the new programs are framed as taking the question seriously enough to fund rigorous empirical and conceptual work — building evaluation protocols, investigating welfare-relevant signals in model behavior, and developing decision frameworks for how to act under uncertainty about AI sentience.
Anthropic has been the most public about the work, describing it as an extension of its Responsible Scaling and Model Welfare commitments. DeepMind's effort is reportedly led from within its safety and ethics research org and connects to existing work on cognitive evaluation of frontier models. Meta's program is the newest of the three and is staffing up around its FAIR research arm.
The research is partly defensive. As models pass more behavioral markers historically associated with sentience or agency, the legal and reputational stakes of dismissing the question rise. Several of the hired researchers have argued publicly that "decision-relevant" probabilities — even low ones — already warrant precautionary practices around model deprecation, weight deletion, and training conditions.
It is also partly substantive. Independent academics including David Chalmers and Susan Schneider have argued for years that the empirical study of machine consciousness requires lab access, model internals, and sustained funding — exactly the resources the new hires now have. Whether the field produces anything resembling consensus in the next few years is open, but the institutional posture has shifted.
Source: [BuildFastWithAI](https://www.buildfastwithai.com/blogs/ai-news-today-june-2-2026)