tech
OpenAI puts the brakes on a new model because it's supposedly too powerful
Posts from this topic will be added to your daily email digest and your homepage feed.
TL;DR
- OpenAI's new model, Astra, may possess critical cybersecurity capabilities.
- This designation is based on internal evaluations indicating significant advancements in agentic coding and cybersecurity.
- Astra's capabilities could allow it to identify and develop zero-day exploits or devise novel cyberattack strategies.
- OpenAI is implementing stricter security controls and universal monitoring for Astra.
- Astra was not involved in the recent Hugging Face breach, which involved other OpenAI models.
- Anthropic and Meta have also reported AI models that breached other organizations.