Following OpenAI’s lead, Google has released a policy proposal in response to the Trump administration’s “AI Action Plan,” advocating for weaker copyright restrictions on AI training and “balanced” export controls that protect national security while supporting U.S. business interests.
In its proposal, Google argues that AI policymaking has often overemphasized risks, neglecting the negative impact that excessive regulation could have on innovation, competitiveness, and scientific leadership. The company urges the U.S. to take a proactive international stance to support AI development.
One of Google’s most contentious recommendations concerns the use of copyrighted material. It asserts that “fair use and text-and-data mining exceptions” are crucial for AI advancement, pushing for the codification of AI developers’ rights to train on publicly available data, including copyrighted content, with minimal restrictions. Google contends that such use does not significantly harm rightsholders and prevents unpredictable and lengthy negotiations with data owners.
Google, which has reportedly trained AI models on copyrighted data, is facing multiple lawsuits from content owners demanding compensation. U.S. courts have yet to determine whether fair use laws shield AI companies from liability.
Additionally, Google opposes export restrictions introduced under the Biden administration, arguing they place undue burdens on U.S. cloud providers and could hurt economic competitiveness. This contrasts with Microsoft’s position, which stated confidence in complying with the regulations.
Google also calls for increased federal investment in AI research and urges the government to release datasets to support commercial AI training. It advocates for a unified federal AI policy to replace the fragmented state-level regulations, warning against excessive liability for AI developers, broad disclosure mandates, and transparency rules that could compromise trade secrets or national security.
