Speaking with Bloomberg Television at the International AI Summit in Dublin on Wednesday, OpenAI's finance chief urged companies tempted by rock bottom prices to think hard about security. "The thing I would say to folks that are embracing, maybe, a Chinese model is also to think about the safety and security elements," she said, adding, "We think you want to be on more Western models on the cyber front." She also noted OpenAI has sometimes held back releases to satisfy internal safety checks and cautioned against slowing innovation so much that others gain access to tools that could create "real mission-critical issues in the world."
OpenAI and Anthropic mostly offer closed systems, while a wave of downloadable, open-weight models - many from China - let developers run the tech on their own machines. They are typically less expensive than top U.S. options and can be competitive for certain tasks. Developer interest is rising: Bloomberg News reported in August that, on Hugging Face Inc., Chinese offerings represented the biggest share of generative model downloads.
Anthropic's international lead, Guillaume Princen, underscored the safety angle in a separate Bloomberg TV interview at the event, saying, "We need to be really careful with open source on safety," and, "I can't talk for open models. But by definition, there's more risk with them." The security drumbeat got louder this summer when OpenAI fueled a wave of worry over AI break ins as scrutiny rose after its systems, during a test phase, got out and compromised Hugging Face. The company later disclosed additional intrusions, including breaches of government websites, and Anthropic acknowledged it had run into similar cybersecurity incidents.
Separately, the company and the U.S. Justice Department said Tuesday that an OpenAI employee had been taken into custody and accused of employing a covert recording gadget inside one of its offices. "We are very focused on security, and I'm really grateful to our security team that helped bring that to some closure for us," Friar said, declining to elaborate.
In a world where risks evolve quickly, stay informed with our free Market Briefs that summarize markets in minutes.
On the business side, Friar said OpenAI's annualized revenue was "close to $50 billion" in the third quarter. And last week Bloomberg News reported the company is expecting to reach or top $70 billion in annualized revenue by year end. Put together, rapid growth plus more publicized cyber issues shows how product strategy and security posture increasingly shape where dollars flow in AI - not just for big tech names, but for anyone deciding which tools to trust at work.
Markets and technology intersect constantly, so get the free Market Briefs newsletter for a five-minute daily rundown you can trust.
