Trump is set to unveil his AI roadmap: Here’s what to know

U.S. President Donald Trump is expected to unveil his long-awaited AI Action Plan at a Washington D.C. event Wednesday hosted by Silicon Valley insiders — his first major address concerning artificial intelligence since he took office for the second time in January.

The AI Action Plan should provide a roadmap of the Trump administration’s strategies, priorities, and concerns around AI — likely a technology that will come to define the 47th President’s term.

The plan is effectively a replacement for the Biden AI executive order, the previous administration’s AI strategy which placed a large focus on mandating AI companies to submit safety and security reports, and trying to limit racial or otherwise discriminatory bias in frontier AI models. Trump repealed Biden’s order within days of his inauguration, arguing that its requirements could be onerous for AI companies, and may hinder American innovation.

In its first six months, the Trump Administration has broadly encouraged efforts to accelerate the development and distribution of American AI technology. Trump helped OpenAI, Oracle, and SoftBank announce their multi-billion dollar Stargate data center project, and the President peeled back restrictions on Nvidia selling its AI chips around the globe.

At the same time, Trump’s AI czar David Sacks has picked a fight with technology companies over “woke” AI, claiming that OpenAI, Anthropic, and Google are supposedly instilling left-leaning values into their AI chatbots and censoring conservative viewpoints.

Some are already pushing back on Trump’s AI Action plan for allegedly putting corporate interests ahead of the public. On Tuesday, a group of more than 90 organizations including labor, environmental justice, and consumer protection non-profits published an open letter called the People’s AI Action Plan. This puts forth a series of AI policies that claim to put the interests of American citizens first, and counter what Trump is expected to announce.

“We can’t let Big Tech and Big Oil lobbyists write the rules for AI and our economy at the expense of our freedom and equality, workers and families’ well-being,” the group said in a statement to TechCrunch that acknowledged the energy needs of Silicon Valley’s AI data centers.

Trump’s AI Action Plan should advance his administration’s agenda more explicitly, but exactly how remains unclear. Trump is expected to share more details about the plan at the “Winning the AI Race” summit, an event hosted by the Hill and Valley Forum and the All In podcast, which Sacks co-hosts when he’s not serving as a government official or venture capitalist.

Here’s what we know about the AI Action Plan so far.

Accelerating American AI

Trump’s AI strategy is expected to focus on three pillars — infrastructure, innovation, and global influence — according to a report from Time Magazine.

For infrastructure, the Trump administration is reportedly planning to overhaul permitting rules to speed up the development of AI data centers. This aims to help AI companies meet with the growing energy needs to train and serve their AI models. However, it’s widely expected that the rise of AI data centers — which suck up immense amounts of energy and water from neighboring communities — could cause energy shortages by the end of the decade unless there’s a rapid increase in energy production.

The president’s infrastructure pillar is also expected to include a plan to modernize America’s electrical grid and add new sources of energy to power these data centers, according to Time.

On the innovation front, Trump reportedly plans to use his AI Action Plan to revive the conversation around blocking state AI laws (even though a federal proposition on the issue overwhelmingly failed last month). This is part of an effort to reduce barriers to innovation for American AI companies, but may ultimately block lawmakers from passing meaningful safety and security standards for AI companies.

As for the global influence pillar, Trump is expected to put forth a strategy to advance the adoption of American AI models and chips, not just in the U.S., but around the world. Federal officials were spooked by the rise of the DeepSeek, and other Chinese AI labs such as Qwen and Moonshot AI have since become worthy competitors to OpenAI. Trump wants America’s technology to be the global standard.

To advance its goals, the Trump administration is also expected to sign a series of AI-related executive orders on Wednesday, according to The Washington Post. Some of these orders clear the path for faster data center buildouts, while others encourage the export of American technologies.

Combatting ‘woke’ AI

One of the executive orders Trump is slated to sign Wednesday would crackdown on “woke” AI models, the Wall Street Journal reported earlier this week. The order requires AI companies with federal contracts — which includes OpenAI, xAI, Google, and Anthropic — to ensure their AI models have neutral and unbiased language.

The crackdown on “woke” AI marks the Republican party’s latest attack on Silicon Valley’s historically left-leaning crowd. In past years, Republican investigated social media companies for allegedly altering their algorithms to censor conservative voices. Meta’s Mark Zuckerberg recently capitulated to these allegations, and overhauled Facebook and Instagram’s content moderation to represent more voices.

A key question looming around this executive order is who defines whether AI models are neutral or biased, and how they determine it. Trump has long stated he’s an advocate of free speech, so an executive order setting rules around what an AI model can and can’t say may seem counterintuitive. That said, a Florida judge recently ruled that AI chatbots are not protected by the First Amendment.

In light of this crack down, OpenAI and other AI labs have tried to make their AI chatbots represent more viewpoints. These companies are in the awkward position to generate AI responses that please everyone, while also not spreading extremist viewpoints or misinformation.

Elon Musk, once Trump’s greatest financial backer whose relationship with the president has recently soured, explicitly started xAI to develop an “anti-woke” AI chatbot, Grok, and combat ChatGPT. However, xAI’s effort to create such a chatbot hasn’t gone so well. In recent weeks, xAI was forced to apologize repeatedly when its AI chatbot went on antisemitic rants and consulted Musk’s personal opinions on hot-button issues.

What Silicon Valley and Big Tech want

The White House announced in April it had received more than 10,000 public comments from companies, local governments, and nonprofit organizations regarding Trump’s AI Action Plan.

OpenAI, Google, Meta, and Amazon effectively took the opportunity to submit wishlists of friendly AI policies they’d like to see the Trump administration implement.

Many of America’s leading AI model developers asked Trump to use his AI Action plan to guarantee that training large language models on copyrighted material would be fair use, and should therefore be allowed. 

Such a protection could benefit these companies significantly, as many of them are involved in active lawsuits with copyright owners from the music, film, news, and book industries. These publishers have accused AI companies of illegally training on their copyrighted works to make AI models, potentially devaluing their media in the process.

Meanwhile, Meta asked Trump to protect open AI models — which are freely available to download online. By releasing its Llama models openly, Meta has been able to undercut OpenAI and Google’s closed offerings. However, Anthropic has raised concerns over whether open AI models could leak powerful technology to bad actors, including China.

Other interest groups, including nonprofits such as The Future of Life Institute, used the commenting period to ask the Trump administration to increase investment into AI research efforts outside of commercial entities. The request comes at a time when the Trump administration and DOGE have slashed funding for American universities, many of which have been powerhouses for scientific breakthroughs in recent decades.

It seems unlikely that Trump’s AI action plan will feature the same safety and security reporting standards that the Biden administration sought to impose. However, polls show that most Americans want AI companies to be held to safety standards.

Several state lawmakers are pushing bills that would create safety and security reporting mandates, but they may face opposition from the Trump administration and Republican lawmakers should they contradict Trump’s AI Action Plan..