Safety and security · Anthropic · OpenAI
The AI Evaluator Forum’s AEF-1 standard has started appearing in independent assessments
This story was corrected after publication (Oct. 9, 2026). Details
The standard, from the AI Evaluator Forum founded in December 2025, sets out five pillars: sufficient access and resources, minimized conflicts of interest, analytic autonomy, transparent methods and results, and protection of sensitive information. It has already started appearing in independent assessments covering Anthropic, Google, Meta and OpenAI. The counter-reading arrived the same week: more than a hundred experts signed an open letter saying Anthropic and OpenAI need truly independent safety evaluators.
Sources
- AI Evaluator Forum, “AEF-1: Minimum Operating Conditions for Independent Third Party AI Evaluations”, (aievaluatorforum.org)
- AI Evaluator Forum, “Minimum Conditions for Embedding Evaluators”, open letter, (aievaluatorforum.org)
- International Business Times, “More Than 100 AI Experts Sign a Letter Saying that OpenAI & Anthropic Need Independent AI Safety Evaluators”, (ibtimes.com)
Corrections and updates
- The headline first posted on Instagram said xAI, OpenAI and Anthropic cosigned the AEF-1 standard. The AI Evaluator Forum’s page lists no company signatories, and no source shows the companies signing it; the headline has been corrected.
About this story
This story was posted on Instagram by @jarrus.tech on Sept. 21, 2026.
Spotted an error in this story? [email protected] · Instagram
Short link: thejarrus.com/en/h6-27
This story in Turkish: AI Evaluator Forum’un AEF-1 standardı bağımsız değerlendirmelerde kullanılmaya başladı
On the same topic
Anthropic releases Claude Haiku 5.5 at 90% lower prices
Claude Haiku 5.5, released Oct. 7, costs 90% less than Haiku 4.5 for prompts up to 100,000 tokens. It leads on benchmarks but uses more tokens per task.
OpenAI is rolling out GPT-6 to everyone in ChatGPT
OpenAI began rolling out GPT-6 to everyone in ChatGPT on October 7. Its new Intelligent UI builds answers from text, visuals and interactive elements.
OpenAI’s unreleased model wrote 722 math papers
On October 6, OpenAI published 722 math papers from an unreleased internal model on GitHub. It withdrew three on October 7, leaving 719 in the catalog.