Get the latest news, advances in research, policy work, and education program updates from HAI in your inbox weekly.
Sign Up For Latest News
In the last year, AI models have rapidly improved on public benchmarks measuring their ability to carry out research tasks autonomously. Such performance suggests that models may be capable of automating parts of the AI R&D process itself, thereby accelerating the pace of AI progress.

In the last year, AI models have rapidly improved on public benchmarks measuring their ability to carry out research tasks autonomously. Such performance suggests that models may be capable of automating parts of the AI R&D process itself, thereby accelerating the pace of AI progress.
This talk will cover key insights from the third AI100 report, AI Goes Mainstream: The Scaling Era and the Rise of Generative AI, set for release in November 2026. The report was commissioned by the AI100 standing committee and led by study panel chair Michael Wooldridge.

This talk will cover key insights from the third AI100 report, AI Goes Mainstream: The Scaling Era and the Rise of Generative AI, set for release in November 2026. The report was commissioned by the AI100 standing committee and led by study panel chair Michael Wooldridge.
Sessions run Wednesdays from 4:30–5:30 PM in CoDa E160. Each session features a different Stanford speaker; talk titles are announced by the organizers.

Sessions run Wednesdays from 4:30–5:30 PM in CoDa E160. Each session features a different Stanford speaker; talk titles are announced by the organizers.
An autonomous system is a software agent capable of performing what appear to be actions, with what appears to be a significant degree of independence and autonomy. Autonomous systems, at least at present, are not genuine moral agents: they do not themselves have moral obligations or permissions. But the programming of an autonomous system is uncontroversially subject to moral evaluation. This talk will consider the general question whether the morality of programming an autonomous system to behave in a certain way in a certain situation is reducible to the morality of a human actor, in its place, behaving in that same way in that same situation. It will consider three ways in which this reduction might fail, based on programmers’ ignorance of who their act affects, on the difference between performing one and multiple instances of the same act, and on the extrinsic effects that can arise from the visibility of programs to others.
