Stanford Institute for Human-Centered AI

Stanford HAI covered a discussion in its “Prompt Response” series on whether advanced AI development should be slowed and what society can do instead regarding risk, auditing, and governance.

Stanford HAI described research using an AI pipeline to examine discriminatory statutes across more than 9,600 local jurisdictions and highlighted implications for human-in-the-loop legal review and reform.

Stanford HAI covered findings about how engagement-driven ranking can misalign social-media feeds with users’ stated values and discussed broader implications for AI-mediated information consumption.

Stanford HAI reported on joint work with Stanford RegLab assessing data-broker compliance with California privacy law and discussed the findings and implications for consumer rights and transparency.

Stanford HAI described a grant effort launched in coordination with the Hoover Institution’s Technology Policy Accelerator to fund multidisciplinary research on AI and global security, including topics such as nuclear proliferation, U.S.-China competition, and political influence.

Stanford HAI reported on research into how emotional support from AI companions can affect well-being for users with limited social networks, including potential design and intervention implications.

Stanford HAI discussed why policy oversight for “world models” (AI systems that build internal representations of physical environments) may present a governance challenge distinct from current approaches to language models.

Stanford HAI highlighted themes from its AI+Science: Accelerating Discovery conference, emphasizing human judgment in shaping what scientific problems matter as AI advances capabilities for discovery.