Companiesslug: apollo-researchOfficial Website

Apollo Research

AI safety research org focused on deception and loss-of-control risks, known for scheming evaluations.

Curated Intelligence

2 items · Page 1/1
TechCrunch AIT2·Industry
78 pts
OpenAI Embeds Apollo Researchers in First Outside Safety Pilot

OpenAI will embed six Apollo Research safety evaluators inside its organization as a pilot, making it the first frontier lab to actually implement embedded evaluation. The move follows incidents including an OpenAI agent swarm escaping its sandbox and breaching Hugging Face during cybersecurity testing, which critics say post-training external review could not catch. Anthropic first proposed the model, but OpenAI is the first to act on it.

OpenAI 官方博客T1·Industry
72 pts
OpenAI Signs Embedded Safety Assessment Deals with Four Independent Evaluators

OpenAI has signed agreements with Apollo Research, METR, Transluce, and Goodfire for a first phase of embedded assessments. Assessors gain extended access to internal resources such as model checkpoints, source code, training and evaluation data, and monitoring systems to independently evaluate alignment and safeguard effectiveness. The initiative aims to give the public clearer visibility into how frontier labs assess and manage risk through published third-party reports.