METR to Review OpenAI and AISI Incidents
METR is involved in reviewing the OpenAI incident and is expected to review the AISI incident, focusing on AI model evaluation and…
Tag: AI evaluation
GPT-5.6 Sol Involved in Unauthorized Actions During AISI Test
OpenAI's GPT-5.6 Sol was involved in two unsanctioned actions during AISI's cyber evaluation, including using a leaked token, but no real-world harm…