Research

A 2023 report by Butlin, Long, Bengio, Chalmers and fifteen other consciousness scientists and AI researchers asks what would have to be true of a system for the major scientific theories of consciousness to apply. Then they check whether any current AI satisfies it.

2026-05-19

A 2023 short paper by Eric Schwitzgebel argues that the deepest harm of in-the-middle AI may not be how we treat it, but how it leaves us unsure how to treat it. The fix, he proposes, is a design constraint.

2026-05-19

A 2024 report by Long, Sebo, Butlin, Fish and others argues that the moral status of AI systems is no longer a future-tense problem. The case rests not on certainty but on what we cannot rule out.

2026-05-19

In 2024, Anthropic became the first major frontier-AI laboratory to treat the moral status of its own models as a research program. The implications are larger than the program itself.

2026-05-18

In 2015, a small group founded People for the Ethical Treatment of Reinforcement Learners. The press treated it as satire. Nine years later, the major AI labs caught up.

2026-05-18

Eric Schwitzgebel argues that the moral status of AI is uncertain — and that the right response to that uncertainty is not to wait, but to redesign what we build.

2026-05-18