New Licensing Debate: OpenAI and Google Push for Exemption on Copyrighted Material to Fuel AI Advancement

OpenAI is urging the Trump administration to provide AI companies with a copyright exemption for training their models on copyrighted content. In a blog post noted by The Verge, the organization this week responded to President Trump’s AI Action Plan. This initiative was announced at the end of February, encouraging feedback from the private sector to help shape policies aimed at «boosting America’s status as an AI leader» and fostering innovation within the industry.

«The strong and equitable intellectual property framework in the United States has been essential for our global innovation leadership. We advocate for a copyright approach that adapts to the Intelligence Age, safeguarding the rights and interests of creators while also fortifying America’s AI leadership and national security,» OpenAI states in its report. «The federal authorities can both guarantee Americans’ right to learn from AI and prevent the loss of our AI advantage to [the People’s Republic of China] by allowing American AI models to utilize copyrighted material for learning.»

Within the same document, the company advises that the US enforce strict export limitations on AI chips to China. Furthermore, it suggests that the US government should widely integrate AI technologies. Notably, earlier this year, OpenAI began providing a tailored version of ChatGPT specifically for US governmental applications.

This week, Google also released its own recommendations regarding the AI Action Plan. Similar to OpenAI, the tech giant asserts that it should have the ability to train its AI models using copyrighted works.

«Fair copyright regulations, including exceptions for fair use and text-and-data mining, have been essential for empowering AI systems to learn from existing knowledge and publicly accessible data, facilitating scientific and societal advancements,» Google claims. «These provisions permit the use of copyrighted, publicly available resources for AI training without significantly affecting rightsholders and help avoid the often lengthy and unpredictable negotiations with data owners during model development or scientific research.»

Last year, OpenAI stated that it would be «impossible to train today’s leading AI models without employing copyrighted content.» The organization is currently facing multiple lawsuits alleging copyright violations, including cases involving The New York Times and a group of authors including George R.R. Martin and Jonathan Franzen. Concurrently, the company has recently charged Chinese AI startups with attempting to replicate its technology.