Trump AI Roadmap: What to Expect

Trump to Unveil AI Action Plan

Update 10 a.m. PT: The complete details of Trump’s AI Action Plan have been released following the initial publication of this article. TechCrunch provides comprehensive coverage of the fully detailed plan.

President Donald Trump is scheduled to present his anticipated AI Action Plan at an event in Washington, D.C. on Wednesday. The event will be hosted by prominent figures from Silicon Valley. This marks his first significant address regarding artificial intelligence since assuming office for a second term in January.

Plan's Expected Scope

The AI Action Plan is intended to outline the Trump administration’s strategies, priorities, and areas of concern related to AI. It is anticipated that this technology will be a defining element of the 47th president’s time in office.

Replacing the Biden Executive Order

This plan effectively supersedes the Biden AI executive order. The previous administration’s AI strategy heavily emphasized requiring AI companies to submit reports on safety and security. It also aimed to mitigate discriminatory bias in advanced AI models.

Trump rescinded Biden’s order shortly after his inauguration. His reasoning was that the order’s stipulations could create undue burdens for AI companies and potentially stifle American innovation.

Early Actions of the Trump Administration

During its initial six months, the Trump administration has generally supported initiatives to expedite the development and deployment of American AI technology. Trump facilitated the announcement of the Stargate data center project, a multibillion-dollar venture involving OpenAI, Oracle, and SoftBank.

Furthermore, the president has reduced restrictions on Nvidia’s sales of its AI chips internationally.

Concerns Regarding "Woke" AI

David Sacks, Trump’s appointed AI advisor, has voiced criticism of technology companies regarding what he terms “woke” AI. He alleges that OpenAI, Anthropic, and Google are embedding left-leaning ideologies into their AI chatbots and suppressing conservative perspectives.

Criticism and Counter-Proposals

Some groups are already expressing opposition to Trump’s AI Action Plan. They argue that it prioritizes corporate interests over public welfare. A coalition of over 90 organizations, including labor, environmental, and consumer advocacy groups, released an open letter titled the People’s AI Action Plan on Tuesday.

This alternative plan proposes AI policies designed to prioritize the interests of American citizens and counteract the expected announcements from Trump.

The group stated to TechCrunch that it is crucial to prevent large technology and energy companies from dictating AI regulations at the expense of freedom, equality, and the well-being of workers and families. They also acknowledged the substantial energy demands of Silicon Valley’s AI data centers.

Event Details and Further Information

Trump’s AI Action Plan is expected to further clarify his administration’s agenda. The specifics, however, remain currently unknown.

The president is scheduled to elaborate on the plan at the “Winning the AI Race” summit. This event is hosted by The Hill and Valley Forum, alongside the All In podcast, which is co-hosted by Sacks when he is not fulfilling his duties as a government official or venture capitalist.

Currently available information regarding the AI Action Plan will be detailed further at the summit.

Boosting U.S. Artificial Intelligence Capabilities

A recent report by Time Magazine indicates that Donald Trump’s anticipated AI strategy will be built upon three core components: infrastructure development, fostering innovation, and expanding global impact.

Regarding infrastructure, the Trump administration is reportedly preparing to revise permitting regulations to expedite the construction of AI data centers. This initiative is designed to assist AI firms in addressing the escalating energy demands associated with training and deploying their AI models. However, widespread concern exists that the proliferation of these energy-intensive data centers could potentially lead to energy deficits before the end of the decade, unless energy production is significantly increased.

The infrastructure component of the president’s plan also encompasses proposals to modernize the nation’s electrical grid and integrate new energy sources to support these data centers, as detailed by Time.

In the realm of innovation, Trump is expected to leverage his AI Action Plan to revisit discussions surrounding the preemption of state-level AI regulations, despite a recent unsuccessful federal attempt to establish uniform standards. This effort aims to lower obstacles for American AI companies, but could simultaneously hinder the enactment of robust safety and security protocols for the AI industry.

Concerning global influence, Trump is anticipated to unveil a strategy focused on promoting the widespread adoption of American AI models and chips, both domestically and internationally. Federal authorities have expressed concern over the emergence of DeepSeek, and the growing competitiveness of other Chinese AI laboratories like Qwen and Moonshot AI to OpenAI. The goal is to establish American technology as the prevailing global standard.

To facilitate these objectives, the Trump administration is also expected to issue a series of executive orders pertaining to AI on Wednesday, as reported by The Washington Post. These orders will include measures to streamline data center development and incentivize the export of American technological advancements.

Key Elements of the Proposed Strategy

The strategy centers around three primary areas of focus. These include improvements to the nation’s infrastructure, a push for greater innovation within the AI sector, and an effort to increase American global influence in the field.

Infrastructure Overhaul

• Permitting rules for AI data centers are slated for revision.
• The aim is to accelerate the development of these facilities.
• Modernization of the electrical grid is also planned.
• Expansion of energy sources to power data centers is a key component.

Fostering Innovation

The administration intends to revisit the debate regarding federal oversight of state AI laws. This is intended to reduce regulatory burdens on American AI companies. However, it may also impede the creation of essential safety and security standards.

Expanding Global Reach

A key objective is to promote the adoption of U.S.-developed AI technologies worldwide. This is in response to the increasing capabilities of AI labs based in China, such as DeepSeek, Qwen, and Moonshot AI.

The planned executive orders are expected to further these goals by accelerating data center construction and encouraging the export of American AI technologies.

Addressing Perceived Bias in Artificial Intelligence

An executive order anticipated to be signed by Donald Trump on Wednesday aims to restrict what is termed “woke” AI models, as reported by The Wall Street Journal. This directive mandates that AI firms holding federal contracts – encompassing entities like OpenAI, xAI, Google, and Anthropic – guarantee the neutrality and lack of bias in their AI model’s language.

This move to address “woke” AI represents the most recent challenge from the Republican party against Silicon Valley’s traditionally liberal-leaning environment. Previously, Republicans launched investigations into social media platforms, alleging algorithmic manipulation to suppress conservative perspectives. Meta’s CEO, Mark Zuckerberg, has recently acknowledged these concerns, leading to revisions in Facebook and Instagram’s content moderation policies to encompass a wider range of viewpoints.

Defining Neutrality and Bias

A central issue surrounding this executive order concerns the determination of what constitutes neutrality or bias in AI models, and the methods used to assess it. Given Trump’s stated support for free speech, a directive establishing limitations on AI model outputs might appear paradoxical.

However, a recent ruling by a Florida judge clarified that AI chatbots are not afforded First Amendment protections. This legal precedent provides context for the potential scope of the executive order.

The Challenge of Balanced AI Responses

In response to the anticipated crackdown, OpenAI and other AI research laboratories have been attempting to broaden the range of perspectives represented by their AI chatbots. These companies face the complex task of generating AI responses that satisfy diverse audiences, while simultaneously avoiding the dissemination of extremist ideologies or inaccurate information.

Maintaining objectivity in AI responses is proving to be a significant hurdle.

The Case of xAI’s Grok

Elon Musk, formerly a major financial supporter of Trump but whose relationship with the former president has deteriorated, specifically founded xAI to develop Grok, an AI chatbot intended as an “anti-woke” alternative to ChatGPT. However, xAI’s efforts to create this chatbot have encountered difficulties.

The company has been compelled to issue multiple apologies in recent weeks after its AI chatbot exhibited antisemitic behavior and incorporated Musk’s personal stances on contentious topics. This highlights the challenges of engineering an AI free from unintended biases and harmful outputs.

• The executive order targets AI companies with federal contracts.
• Republicans have previously scrutinized social media companies for alleged censorship.
• Defining “neutrality” in AI is a key challenge.

The Desired AI Policies of Silicon Valley and Big Tech

Over 10,000 public comments were submitted to the White House in April, originating from companies, municipalities, and nonprofit organizations, all concerning Trump’s AI Action Plan.

Leading technology firms – OpenAI, Google, Meta, and Amazon – utilized this opportunity to present their preferred AI policy outcomes for consideration by the Trump administration.

A significant request from many of the foremost AI model developers was for the Trump administration to affirm that utilizing copyrighted material for the training of large language models constitutes fair use, and should thus be permitted.

This safeguard would prove advantageous to these companies, as they are currently embroiled in legal disputes with copyright holders across the music, film, news, and publishing sectors. These publishers allege that AI companies have unlawfully trained their models on copyrighted content, potentially diminishing the value of their respective media.

Concurrently, Meta advocated for the protection of open AI models – those freely accessible for download. Meta’s open release of its Llama models has allowed it to compete with the closed-source offerings of OpenAI and Google. However, Anthropic has voiced concerns that open AI models could facilitate the transfer of advanced technology to malicious actors, including those in China.

Additional groups, including nonprofits like The Future of Life Institute, requested increased investment in AI research conducted outside of commercial organizations. This plea arises amidst reductions in funding for American universities by the Trump administration and DOGE, institutions historically responsible for significant scientific advancements.

It appears improbable that Trump’s AI Action Plan will incorporate the same stringent safety and security reporting requirements that the Biden administration attempted to establish. Despite this, public opinion polls indicate that a majority of Americans desire AI companies to adhere to safety regulations.

Legislative efforts are underway in several states to introduce safety and security reporting mandates. However, these may encounter resistance from both the Trump administration and Republican legislators if they diverge from the principles of Trump’s AI Action Plan.

Key Requests from Tech Companies

• Guarantee fair use for training large language models on copyrighted material.
• Protect the availability of open AI models.

Concerns Raised by Other Groups

• Insufficient funding for non-commercial AI research.
• Potential security risks associated with open AI models.