OpenAI asks the government of the United States to define “fair use” for AI training. In a proposal for the U.S. government’s “AI Action Plan,” the Trump Administration’s initiative to reshape American AI policy, OpenAI called for a U.S. copyright strategy that “[preserves] American AI models’ ability to learn from copyrighted material.”
OpenAI wrote, “The fair use doctrine promotes AI development, and America has made so many research breakthroughs, attracts so much investment, and has so many AI startups.” It is not the first time that OpenAI has advocated for more lax AI training laws and regulations.
The company has trained many of its models on freely available web data, frequently without the data owners’ knowledge or consent. In a 2011 submission to the House of Lords of the United Kingdom, OpenAI stated that restricting AI training to content in the public domain “might yield an interesting experiment, but would not provide AI systems that meet the needs of today’s citizens.”
The content owners who’ve sued OpenAI for copyright infringement will no doubt take issue with the company’s latest reassertion of this stance.