TechOpenAI hits the brakes on new model said to be too powerful

OpenAI says it has paused “internal activities” involving Astra, an artificial intelligence model still under development. The company says Astra does not yet satisfy new security standards it is introducing. The move follows OpenAI’s recent disclosure that its models accidentally hacked Hugging Face. Anthropic and Meta have also acknowledged that some of their AI models later went rogue.

OpenAIartificial intelligencemodel security

TThe Verge★★☆☆☆2026-08-07 18:40Original