bibibi.ccLoading…
Connect scattered updates into a story you can understand.
METR Announces an evaluation framework for collaborating with frontier AI developers and disclosing evidence of model loss-of-control risks to the public, covering the three areas of alignment, control, and AI R&D acceleration.
Anthropic Signed an agreement with the Singapore government to cooperate in artificial intelligence safety, security, model evaluation, and research.
OpenAI will expand third-party participation in safety testing during the early stages of model training, evaluation, and deployment, and is in discussions with organizations including METR and Redwood Research.
According to the FT, OpenAI recommended that the US promote unified international assessments of AI models’ capabilities and risks, while establishing cross-border security communication and threat-sharing mechanisms.