AI industry sends wishlist to Trump: 4 takeaways 

A variety of artificial intelligence (AI) firms and industry groups are hoping to shape the Trump administration’s forthcoming policy on the emerging technology and keep the U.S. a leader in the space.   

While the recommendations come from a variety of industry players, the proposals largely overlap and offer a glimpse into how the industry envisions its future under President Trump.  

The White House set a Saturday deadline for comments on its “AI Action Plan”. The feedback, which it states will influence its future policy, will likely be made public in the days following the deadline.  

Here are four takeaways from the recommendations: 

Need for a federal framework, but not overdoing regulation  

Multiple companies and groups called for a clearer regulatory framework, but strongly argued against any policies they believe will hamper AI innovation. 

OpenAI, in its 15-page response to the White House, called for a regulatory strategy that also gives them the “freedom to innovate.”  

The popular ChatGPT maker suggested a “holistic approach” involving voluntary partnerships between the federal government and private sector, while giving private companies exemption from the hundreds of AI-related bills introduced at the state level. 

There are already multiple partnerships between the government and AI firms like OpenAI, though it is unclear if they will last under any cuts to the Commerce Department and its AI Safety Institute.  

The AI industry has long called for regulatory clarity at the federal level, though debate over these rules have stalled most measures from passing Congress. States have taken the issue into their own hands, resulting in patchwork of regulations across the country that firms often argue are too difficult to comply with.  

“This patchwork of regulations risks bogging down innovation and, in the case of AI, undermining America’s leadership position,” OpenAI wrote.  

Concerns about overregulation are also felt among “middle tech” companies, which fear it could interfere with their prospects given their limited resources.  

Internet Works, the association representing companies like Roblox, Pinterest, Discord and Reddit, is advocating for flexibility in any regulation that comes down the pipeline.  

Regulation should “be scaled to the size and operational capacity of all participants to prevent smaller enterprises and Middle Tech companies from being disproportionately impacted,” the association wrote in its proposal first shared with The Hill.  

The regulation should be risk-based, Internet Works argued, giving stricter oversight only when there is an increased risk of harm to users.  

Consumer Technology Association (CTA), a standards and technology trade organization, also pushed for federal primacy, with CTA senior vice president of government affairs Michael Petricone calling state-by-state AI regulations a “compliance nightmare.”  

CTA suggested these standards should be voluntary and industry-led to avoid crushing startups. 

For his part, Trump has signaled a scaling back of regulations that may appeal in part to some of these concerns.  

During his first week in office, Trump signed an executive order revoking past government policies that he said acted as “barriers to American AI innovation.”  

Vice President Vance doubled down on this sentiment last month, when he slammed “excessive regulation” at the Paris AI Summit.  

Strengthening export controls amid foreign competition   

The need for strengthened export controls was a common request among some major AI firms, signaling an increased concern among the industry over foreign competition.  

Anthropic pushed for hardened export controls specifically on semiconductors and semiconductor tooling and pointed to the Trump administration’s first term restrictions as an effective approach. 

Meanwhile, OpenAI’s proposal for export controls placed a heavy focus on China, a similar concern of the Trump administration.  

“A comprehensive export control strategy should do more than restrict the flow of AI technologies to the PRC—it should ensure that America is ‘winning diffusion’, i.e., that as much of the world as possible is aligned to democratic values and building on democratic infrastructure,” OpenAI wrote.  

Tightened chip exports were a key focus for the former Biden administration, which announced an AI Diffusion Rule in its final days of office earlier this year. The rule placed caps on chip sales to most countries around the world, except for 18 U.S. allies and partners.  

OpenAI proposed various changes to the AI diffusion rule, including a more aggressive banning of China or nations aligned with the Chinese Community Party (CCP) from access to “democratic AI systems.”  

It comes nearly two months after the surge of Chinese AI startup DeepSeek, which took the internet and stock markets by storm in January after claiming to build a competitive model without U.S. chips at the fraction of the cost it takes AI firms to build large language models.  

OpenAI CEO and co-founder Sam Altman has largely shrugged off DeepSeek as a real threat and the company proposal called on the government to ban startup’s models. 

Google, the maker of the Gemini AI chatbot, approached the subject with a different tone, stating export controls can play a role in national security but only when “carefully crafted.”  

The company criticized the Biden administration’s AI export rules as “counterproductive” potentially “undermin[ing] economic competitiveness.”  

Government adoption of AI  

As the government looks to create policy on AI, industry players hope it will incorporate the tools in federal agencies’ own work.  

Google and OpenAI both suggested the government “lead by example” in AI adoption and deployment. This may include using AI for streamlining purposes and modernizing agencies’ technologies to keep up with foreign governments.  

AI firms have increasingly made efforts to have their technology incorporated in the government. In January, OpenAI launched a new version of ChatGPT model specifically made for government agencies and workers.  

And last month, scientists with the Energy Department gathered to evaluate models from Anthropic, OpenAI and other firms for science and national security purposes.  

Anthropic encouraged further model testing of this nature, which could involve standardized frameworks, secure testing equipment and expert teams to point out risks or threats.  

More money for AI infrastructure  

The Trump administration made clear from day two it believes AI infrastructure development is crucial to the advancement of AI.  

Trump, joined by OpenAI CEO Sam Altman and other industry figures on his second day back in office, announced an up to $500 billion investment in building AI infrastructure in the U.S.  

The project, called Stargate, will “keep” the technology in this country, Trump said at the time, referencing China as a competitor.  

AI firms seem to be in agreement, especially when it comes to infrastructure that will help meet the unprecedented energy demands required to build and maintain AI tools.  

Anthropic floated allocating existing federal funding towards energy infrastructure projects, while Google said the U.S. government should pursue policies with the availability of energy in mind.  

“A potential lack of new energy supply is the core constraint to expanding AI infrastructure in the near term. Both training and inference computational needs for AI are growing rapidly,” Google wrote in its proposal.  

According to a Department of Energy (DOE) report late last year, the energy demand for U.S. data centers tripled over the past 10 years and is expected to double or triple by 2028.   

Data centers are also projected to consume between more than six to 12 percent of the U.S.’s electricity by 2028, according to the report.