✦ For YouGeopoliticsTechFinanceHealthEnergySportsCulture◆ SN Last Week★ Saved
📎 SkimNews has covered OpenAI 460+ times · see the file →

Anthropic, OpenAI Pledge Embedded AI Safety Evaluators — SkimNews

By TechCrunch · Summarized & edited by · 2026-09-16
Anthropic, OpenAI Pledge Embedded AI Safety Evaluators

Get the Tech newsletter

Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.

Why it matters: The proposal's credibility hinges on whether Anthropic and OpenAI actually surrender control — Far.AI's history of rejecting restrictive contracts and Apollo's three-day Astra window suggest frontier labs default to limiting evaluator independence, and without legislation (which Papadatos says is the only durable safeguard), voluntary commitments could evaporate under commercial pressure.

Share this story

Ask SkimNews
More tech → Read original →

Get the Tech newsletter

Curated tech stories, every morning. Free.

No spam. Unsubscribe anytime.