Why did OpenAI slow Astra's development?
OpenAI found preliminary evidence that Astra may meet the Critical cybersecurity capability threshold in its own Preparedness Framework. That threshold means the model could independently identify and execute cyberattacks against traditionally well-protected systems, prompting the company to slow certain aspects of development.
Answered in
OpenAI Slowed Astra Development After Hitting Its Own Cybersecurity Safety ThresholdThe same model solving century-old math problems triggered OpenAI's own safety framework, prompting it to slow development over critical cybersecurity concerns.
Read the full analysisOther questions this article answers
More ai & society questions
- Is model distillation always an attack?
- How much did Anthropic say its new controls reduced extraction?
- Does Anthropic's self-assessment block harmful requests?
- How accurate is Claude's self-assessment monitor?
- Are ChatGPT locked chats end-to-end encrypted?
- Do locked chats change model training or retention?
- What did the dissertation study measure?
- Does an AI-detector flag prove a student used AI improperly?
Every answer on Crashtech is written by the editor of the article it comes from — never auto-summarised. Browse all answers or the AI & Society beat.