Google LLC has submitted a set of suggestions for how the U.S. government can support the domestic artificial intelligence ecosystem.

The search giant outlined its recommendations in a 12-page paper published on Thursday. The same day, OpenAI submitted a similar policy document to the White House.

The two companies drafted the recommendations in response to an AI-focused executive order that President Donald Trump signed earlier this year. The order calls for the creation of a policy framework, dubbed the AI Action Plan, that will determine how the administration approaches the AI market. In February, the White House asked tech firms and other interested parties to submit ideas for the policy framework.

Google’s 12-page paper contains more than a half-dozen recommendations. Several of the suggestions focus on AI-related legislative matters, most notably the way copyright law applies to training data.

Google’s paper backs copyright exceptions that “allow for the use of copyrighted, publicly available material for AI training.” The search giant argues that such exceptions are important for, among other things, AI tools that speed up scientific research. In particular, Google points to machine learning applications that allow researchers to sift quickly through a large number of academic papers.

Another legislative idea the company floated relates to state-level AI laws. Google is suggesting the White House “support federal legislation that prevents a patchwork of laws at the state level, especially for frontier AI development.” The company argues that a fragmented regulatory landscape could slow AI development.

Yet another set of policy recommendations in Google’s paper focuses on the federal government’s day-to-day operations. Agencies “should lead by example in AI adoption” by using machine learning products from multiple suppliers, Google argues. According to the company, the U.S. Commerce Department bureau tasked with enforcing export controls could use AI for supply chain tracking. Google’s paper goes on to argue that the bureau should “avoid creating undue competitive disadvantages” for U.S. AI providers.

In the utility sector, Google believes policy changes are needed to meet the electricity demands of AI data centers. The company argues that the permitting process for new energy infrastructure should be streamlined. Additionally, it sees a need for incentives that would encourage utilities to use their existing infrastructure more efficiently.

Another section of Google’s paper focuses on AI research. According to the company, the U.S. government should direct more funding to “early-market R&D” and support federal prize challenges that encourage AI breakthroughs. Google also sees a need to make technical resources such as compute infrastructure more accessible for researchers.

There are some similarities between the search giant’s suggestions and the policy paper that OpenAI submitted for the White House’s AI Action Plan. Like Google, the ChatGPT rival is reportedly “seeking some relief” from state-level AI rules. OpenAI is also asking officials to relax the copyright rules that apply to data used in AI training projects. 

Photo: Unsplash