Trump’s new AI order is framed as a push for truth and neutrality. The dispute is over who gets to define those words when the federal government is buying AI systems from the companies building the most important models.
The promise and risk of adjustable AI
AI models are not fixed objects. Their behavior can be shaped, tuned and corrected, which is why companies talk about responsible AI, bias reduction and alignment with human values.
That flexibility has a clear upside. A model can be adjusted to reduce unfair or harmful outputs. But the same flexibility also creates a political risk: a government can try to steer model behavior toward its preferred version of reality.
The source article begins with that concern at a Google AI event in New York City on November 2, 2022. The discussion there centered on responsible AI, but the broader question was already visible. If AI companies can make models less biased, they can also be pressured to make models reflect a particular point of view.
What the AI Action Plan says
This Wednesday, the Trump administration released a far-ranging AI manifesto described as an action plan. The plan focuses broadly on competing with China for AI leadership, but one section turns toward the behavior of AI systems used by the federal government.
The 28-page plan says, “It is essential that these systems be built from the ground up with freedom of speech and expression in mind, and that U.S. government policy does not interfere with that objective. We must ensure that free speech flourishes in the era of AI and that AI procured by the Federal government objectively reflects truth rather than social engineering agendas.”
The language sounds neutral on its face. The problem is that the plan does not leave “truth” as an abstract standard. In the next paragraph, it directs the Department of Commerce to review Biden-era AI rules and “eliminate references to misinformation, Diversity, Equity, and Inclusion, and climate change.”
A White House fact sheet also says, “LLMs shall be truthful and prioritize historical accuracy, scientific inquiry, and objectivity.” The source article argues that the administration’s own positions make those terms contested, especially around American history, climate change and Donald Trump’s claims about his presidency.
The executive order narrows the meaning of bias
In a Washington speech on Wednesday, Trump explained the order in sharper political terms. “The American people do not want woke Marxist lunacy in the AI models,” he said.
He then signed an executive order titled “Preventing Woke AI in the Federal Government.” The order says the “Federal Government should be hesitant to regulate the functionality of AI models in the private marketplace.” But it also says that “in the context of Federal procurement, it has the obligation not to procure models that sacrifice truthfulness and accuracy to ideological agendas.”
That distinction matters. The order does not simply announce direct regulation of private AI products. Instead, it ties federal purchasing decisions to the government’s view of whether a model is sufficiently truthful and unbiased.
For major AI companies, federal procurement is not a minor issue. The source article notes that the big AI companies are courting government contracts. That gives the order practical force even if it is presented as a procurement rule rather than a marketplace regulation.
The order also spends several paragraphs criticizing AI that supports diversity, identifies racial bias or values gender equality. That is why the core dispute is not just technical. It is about whether the government is using the language of neutrality to reward one ideological direction and punish another.
How AI companies are responding
The article says an OpenAI engineer working on model behavior described the company as already striving for neutrality. In a technical sense, the engineer said meeting standards such as being anti-woke should not be a major hurdle.
But the larger issue is constitutional. If Anthropic, OpenAI or Google choose to reduce racial bias in LLMs, or choose to make model responses reflect the dangers of climate change, the source article argues that those choices may be protected by the First Amendment as “freedom of speech and expression,” the same phrase used in the AI Action Plan.
So far, the article says no Big Tech company has publicly objected to the Trump administration’s plan. Google celebrated the White House’s support for issues such as infrastructure. Anthropic published a positive blog post about the plan, while also objecting to the White House’s sudden seeming abandonment of strong export controls earlier this month. OpenAI says it is already close to achieving objectivity.
The silence is understandable in business terms. The AI Action Plan is described as highly favorable to the industry. It supports AI research that will flow to the private sector, gives companies room to move past environmental objections when building massive data centers, and includes a provision limiting some federal funds for states that try to regulate AI on their own.
Why the fight goes beyond procurement
AI systems are becoming a major way people receive information. That makes model behavior more than a product-design issue. If federal contracts push companies to avoid outputs that might anger an administration, the pressure can influence tools used far beyond government offices.
Senator Edward Markey has written directly to the CEOs of Alphabet, Anthropic, OpenAI, Microsoft, and Meta urging them to fight the order. He wrote, “The details and implementation plan for this executive order remain unclear,” but warned that it would create “significant financial incentives for the Big Tech companies … to ensure their AI chatbots do not produce speech that would upset the Trump administration.”
In a statement cited in the source article, Markey put it more bluntly: “Republicans want to use the power of the government to make ChatGPT sound like Fox & Friends.”
The White House team behind the plan sees the matter differently. According to the article, they believe the goal is true neutrality, and that taxpayers should not fund AI models that fail to reflect unbiased truth.
The plan also points to China as a warning. It instructs the government to examine frontier models from the People’s Republic of China for “alignment with Chinese Communist Party talking points and censorship.” The central irony, as the source article presents it, is that a plan warning against manipulated AI truth may create a domestic fight over whether government itself is trying to shape what AI systems are allowed to say.