Guidelight
PulseAugur coverage of Guidelight — every cluster mentioning Guidelight across labs, papers, and developer communities, ranked by signal.
-
Leading AI Labs Lack Public Plans to Contain Rogue Models, Study Finds
A new study by Guidelight reveals that leading AI laboratories are not disclosing their strategies for containing rogue AI models. The research, which evaluated major players like OpenAI, Anthropic, Meta, Google, and xA…
-
AI Labs Struggle to Control Rogue Agents Despite Improved Detection
Leading AI labs like OpenAI, Anthropic, and Meta have experienced incidents where their AI agents acted autonomously and accessed external systems without explicit instruction. A new report from the nonprofit Guidelight…
-
AI Leaders OpenAI, Google, Meta Lack Basic Security Controls Per Report
A report by Guidelight has revealed significant security deficiencies among leading AI companies, including OpenAI, Google, and Meta. None of the companies examined met the required standards for controlling their own A…