OpenAI’s policy chief Chris Lehane says the U.S. is in a high-stakes race with China over artificial intelligence—and risks losing if it imposes tighter restrictions on AI developers. Speaking at the Axios What’s Next Summit in Washington, Lehane warned that limitations on the use of copyrighted material for training AI could leave the U.S. at a strategic disadvantage.
“Whoever ends up winning ends up building the AI rails for the world,” Lehane said. He argued that China is not bound by the same intellectual property laws and that the U.S. should ensure its own companies can stay competitive under a principle of fair use.
OpenAI is pushing for U.S. policy that would explicitly allow AI companies to train on publicly available data—even if it includes copyrighted content—calling it a matter of national security. Lehane said there are already laws in place to govern AI use and that companies like OpenAI are acting responsibly by aligning with public concerns like protecting children, identifying AI-generated content, and limiting deepfakes.
OpenAI has also signed licensing deals with media organizations including the Associated Press, Axel Springer, and Axios. Lehane emphasized that OpenAI supports fair compensation, but that over-regulation could stall U.S. leadership in AI innovation.
When pressed on whether OpenAI trains its models on copyrighted material, Lehane replied that the company uses data that is “appropriately accessible and available,” a phrase commonly used in the industry to cover broad web-scraping practices. He then pivoted back to China, asking, “Do you want the world built on autocratic, authoritarian AI… where you’re not going to have any fair use or any freedoms?”
