Why OpenAI and Anthropic are risky for different reasons than their Chinese AI rivals
AI researchers are highlighting distinct security risks between American frontier labs like OpenAI and Anthropic versus open-weight models from Chinese competitors. While American labs face risks related to advanced agent capabilities, open-weight models allow users to bypass safety guardrails.
Why it matters
Understanding these different risk profiles is critical for policymakers and developers as AI capabilities rapidly advance globally.
OpenAI CEO Sam Altman. Bloomberg/Getty Images Researchers spent six days inside OpenAI, investigating this summer's Hugging Face hack. Open models get a lot of attention when it comes to AI risks. This hack was a frontier lab problem. Two of the researchers expressed concerns that labs will lose control over cutting-edge AI. OpenAI and Anthropic have become "fertile ground" for AI's riskiest dangers, an AI researcher says. That's what Ajeya Cotra, who investigated OpenAI's security incident with Hugging Face, told Business Insider, as the event sparked a new wave of fears about AI-driven cyberattacks. Besides concerns about OpenAI's models, some in the industry worried about open-weight models from Chinese labs because users can download them and edit away their safety guardrails, and the models aren't far behind OpenAI's.
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