A Quick Guide to the US Executive Orders on AI

The Trump administration sets out its framework for AI

Posted on July 30, 2025 in Blog Posts

This past week, President Trump signed three executive orders solidifying the US’s stake in AI development. This comes after the tortuous passage of the Big, Beautiful Bill in Congress saw the removal of a key provision that prevented states from issuing their own regulations on AI for 10 years. AI companies supported this moratorium because they felt the process of navigating different state regulations would frustrate technological development. Many politicians and commentators, however, worried that unregulated AI had the potential to create immense and yet-unknown problems. Though this provision ultimately didn’t end up in the final bill, its spirit lives on in these executive orders. 

For non-US readers, executive orders are directives signed by the president that guide the federal government’s policies and behaviors. Though they have legal weight, they are not laws themselves and are arguably, due to congressional gridlock, far easier to alter and remove than laws are. (Customarily, when the presidency changes parties, the new president undoes many of the previous president’s executive orders on day one.) Because the US is currently the global pioneer of AI technology, it’s worthwhile for readers outside the US to understand the framework the US is putting in place with these orders. 

Indeed, the first executive order, which we will return to more closely below, makes the US’s global ambitions quite plain: 

Artificial intelligence (AI) is a foundational technology that will define the future of economic growth, national security, and global competitiveness for decades to come. The United States must not only lead in developing general-purpose and frontier AI capabilities, but also ensure that American AI technologies, standards, and governance models are adopted worldwide to strengthen relationships with our allies and secure our continued technological dominance. 

Unspoken is that the US’s interest in AI is primarily a challenge to Chinese technological dominance, but AI development under these executive orders will inevitably exceed geopolitics and have ramifications for many aspects of contemporary life across the globe. 

---

🔥 Stay up-to-date with LibTech Insights by signing up for our free newsletter. Just one weekly email with our new blog posts, top tech news stories, and other bonus content. Check out some posts from our archive:

• Why Your Academic Library Should Be on TikTok
• Which Digital Content Types Should Be on Your Library’s Radar?
• How to Host a Website Cleanup Event

🌟 Register now for our free email micro-course on AI literacy!

---

EO: “Promoting the Export of the American AI Technology Stack” 

A stack refers to the combination of technologies—programming languages, databases, software, protocols, chips, and more—that underlie and run an application. This first executive order, therefore, doesn’t merely announce that the US wants to lead the market, but it wants to create the frameworks that undergird individual products made anywhere on the globe.

The executive order immediately makes clear the foreign policy implications: the US doesn’t want to rely on another country’s stack, which would leave the US’s tech infrastructure economically, politically, and militarily vulnerable to outside forces. It also doesn’t want its allies to have these vulnerabilities or debts if they should choose to use alternative stacks developed by another country (i.e., China). This executive order seeks to beat US competitors to the punch by fast-tracking the export and adoption of its own stack across the globe.

EO: “Accelerating Federal Permitting of Data Center Infrastructure” 

Lost in the initial panic surrounding ChatGPT’s launch that “AI will take everyone’s job” is that no country has the energy or data infrastructure necessary to use AI at such a scale. This executive order focuses on the material infrastructure for AI. It attempts to facilitate the process of building AI-critical data centers. An executive order can’t create infrastructure on its own, but it can ease regulatory processes to fast-track their production. (See, for instance, Operation Warp Speed to develop COVID vaccines.)  

That’s essentially what this executive order does by directing regulators to ease the burdens they place on AI-critical infrastructure, permitting the development of data centers on federally owned lands, and creating incentives, such as loans and loan guarantees, for private companies to build data centers. This executive order will have significant environmental consequences, as it also eases environmental restrictions that would slow or prevent data center development. 

EO: “Preventing Woke AI in the Federal Government” 

The first two executive orders dealt with the technological framework for AI and its material infrastructure. This third order addresses the actual outputs of AI models. As its title suggests, it takes aim at alleged left-wing ideological influences behind AI outputs. It names DEI specifically—including “critical race theory, transgenderism, unconscious bias, intersectionality, and systemic racism”—as a threat to “reliable AI.” Rather bizarrely, it cites an instance where a model generated an image of a female pope as a clear sign that these models “prioritize DEI requirements at the cost of accuracy.”

Again, an executive order is not a law but a directive to federal agencies, and the real force of the order primarily affects government procurement of AI products more than the behaviors of private tech companies. Of course, private companies seeking lucrative government contracts are thus incentivized to comply with the order, so the order will likely have downstream effects on AI products.

Analysis 

While aspects of these executive orders are concerning, the stances they take on China, the environment, federal lands, and DEI aren’t surprising from the Trump administration. These executive orders were accompanied by an AI Action Plan, outlining the US strategy for global AI dominance, centered on innovation, infrastructure, and international security. The Action Plan breaks down the government’s plan to integrate and advance AI into many small areas, but the overall message is clear: the Trump administration plans to move full steam ahead with AI. 

For librarians and information professionals, especially in universities, the third executive order is likely the most relevant. Conceivably, individual states may adopt a similar wording to shape technology procurement decisions, perhaps including at public schools and universities. That’s merely speculation, but I do think it’s worth highlighting how the order defines its two “unbiased AI principles”: 

(a) Truth-seeking. LLMs shall be truthful in responding to user prompts seeking factual information or analysis. LLMs shall prioritize historical accuracy, scientific inquiry, and objectivity, and shall acknowledge uncertainty where reliable information is incomplete or contradictory. 

(b) Ideological Neutrality. LLMs shall be neutral, nonpartisan tools that do not manipulate responses in favor of ideological dogmas such as DEI. Developers shall not intentionally encode partisan or ideological judgments into an LLM’s outputs unless those judgments are prompted by or otherwise readily accessible to the end user. 

While most would agree that LLMs should be truthful, prioritize historical accuracy and scientific inquiry, acknowledge uncertainty, and be nonpartisan, I think the order itself, prefaced by an attack on DEI, suggests that what it calls “neutrality” is explicitly partisan. As we have seen recently in public life, basic matters of “historical accuracy” and “scientific objectivity” have been questioned and dismissed as ideological products rather than regarded as facts in the historical or scientific record. It remains to be seen how these two “unbiased” principles will be reconciled. 

Librarians have long examined the character and tilt of seemingly neutral “information,” which makes them key guides for navigating this landscape. These executive orders reinforce the case that AI literacy will be an important skill for librarians to both cultivate and impart.

Related Posts