In a move echoing OpenAI’s recent initiatives, Google has released a policy proposal in reaction to the Trump Administration’s request for a national “AI Action Plan.” The tech behemoth has voiced support for relatively lenient copyright regulations concerning AI training and proposed “balanced” export controls designed to “safeguard national security while facilitating U.S. exports and global commercial activities.”
Google asserts that “the U.S. must adopt an active international economic policy to promote American values and support global AI innovation.” The company believes that AI policymaking has historically focused excessively on the risks, often neglecting the potential costs of ill-conceived regulations on innovation, national competitiveness, and scientific leadership — a trend they hope is changing with the current Administration.
One of the more contentious suggestions from Google relates to the handling of intellectual property-protected content.
The company contends that “fair use and text-and-data mining exceptions” are “essential” for the progression of AI technology and associated scientific advancements. Similar to OpenAI, Google aims to establish a legal right for itself and its competitors to utilize publicly accessible data — including copyrighted material — predominantly without limitations.
According to Google, “These exceptions facilitate the use of copyrighted material that is publicly available for AI training without significantly harming rightsholders,” and they help to avoid unpredictable, unbalanced, and protracted negotiations with data holders during model creation or scientific research.
The company, which has utilized a variety of publicly available copyrighted materials for training numerous models, is currently experiencing legal challenges with content owners accusing it of failing to inform or compensate them beforehand. U.S. courts have yet to determine if the fair use doctrine adequately protects AI developers from intellectual property lawsuits.
In its policy proposal, Google criticized the export controls implemented by the Biden Administration, claiming they “might hinder economic competitiveness objectives” by “placing undue burdens on U.S. cloud service providers.” This statement contrasts with remarks from competitors such as Microsoft, which recently expressed confidence in its ability to fully comply with the new regulations.
Notably, these export regulations include exemptions for reputable companies that require large quantities of advanced AI chips.
Moreover, Google advocates for “long-term, sustained” funding for foundational domestic research and development, countering federal initiatives aimed at reducing spending and discontinuing grant programs. The company recommends that the government share datasets beneficial for commercial AI applications and invest in “early-market R&D,” ensuring that computing resources and models are “widely accessible” to scientists and research institutions.
Highlighting the tumultuous regulatory landscape stemming from the inconsistent state-level AI legislations in the U.S., Google called for comprehensive federal AI legislation, including a cohesive privacy and security framework. As of early 2025, 781 AI-related bills are pending in the U.S., according to an online tracking resource.
Google warns the U.S. government against instituting what it views as burdensome requirements concerning AI systems, such as liability obligations related to usage. The company argues that in many instances, a model’s developer “has little to no visibility or control” over how a model is ultimately used and therefore should not be held accountable for its misuse.
“Even in scenarios where a developer provides a model directly to end-users, those users often possess the best understanding of the risks associated with its downstream applications, enabling them to implement effective risk management, monitoring, and logging,” Google asserted.
Lastly, Google characterized proposed disclosure requirements similar to those being considered by the EU as “excessively broad,” urging the U.S. government to reject transparency mandates that could “reveal trade secrets, allow competitors to replicate products, or jeopardize national security by offering adversaries a guide to bypassing protections or modifying models.”
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
