Pacing model development in an era of cyber-critical capabilities
An AI research organization is slowing down the development of its latest models to prioritize safety and alignment. This decision follows internal incidents and the identification of potential cybersecurity risks in their upcoming model, Astra.
Why it matters
It underscores the industry-wide tension between rapid AI scaling and the need for rigorous safety protocols to prevent catastrophic outcomes.
Loading… Share Strengthening safeguards for more capable models Strengthening safeguards for more capable models Securing our research environments Expanding chain-of-thought monitoring Advancing alignment research What’s next Strengthening safeguards for more capable models Securing our research environments Expanding chain-of-thought monitoring Advancing alignment research What’s next Over the past several weeks, two developments have underscored the growing risks associated with increasingly capable AI systems: the OpenAI-Hugging Face incident and, separately, preliminary evidence that one of our upcoming models, Astra, may meet the Critical cybersecurity capability threshold under our Preparedness Framework . Together, these developments, combined with rapid progress in our internal research, have added urgency to our work on strengthening our monitoring, alignment, and containment safeguards across all stages of the training process.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in