Last week the White House released an AI Action Plan and executive orders that threaten to withhold federal funds from states that regulate AI, order federal agencies to end investigations into AI companies and delete any existing rules or agreements that could slow AI development, and fast-track or in some cases completely nix the environmental permitting process for AI infrastructure projects.
But there was one string attached to the big tech giveaway: If AI companies want to do business with the federal government, they’d better comply with the Trump administration’s version of the truth.
The order, titled “Preventing Woke AI in the Federal Government,” requires that government agencies only buy AI that is “ideologically neutral” and free of DEI and “partisan or ideological judgements.” It also requires all AI programs purchased by the federal government to “prioritize historical accuracy, scientific inquiry, and objectivity,” and “acknowledge uncertainty where reliable information is incomplete or contradictory.”
On the surface, this seems pretty innocuous. After all, an AI program meant to increase the efficiency of a Veterans Affairs hospital probably shouldn’t be loaded down with political bias and should adhere to scientific accuracy. But this administration and its Republican allies have often worked to create an alternative version of the truth, reject scientific consensus, and weaponize the idea of objectivity to attack news agencies and even big tech companies.
In July, Missouri Attorney General Andrew Bailey provided a perfect example of how this executive order could be weaponized by the government to bully tech companies into complying with their version of the truth. He sent letters to Google, OpenAI, Meta, and Microsoft, threatening them with legal action and demanding to know why their chatbots did not rank President Trump first among the last five presidents in combating antisemitism.
Bailey’s office said the letters were a “continuation” of the attorney general’s “commitment to defending free speech.”
Defending free speech is a common refrain for the current administration when attacking the First Amendment. Federal Communications Commission Chairman Brendan Carr bills himself as a champion of free speech, even while investigating news organizations over editorial decisions such as how they excerpted an interview with a presidential candidate or their choice to describe wrongfully deported Salvadoran immigrant Kilmar Abrego Garcia as a “Maryland man.” Carr also required CBS to hire an ombudsman—which he described as a “bias monitor”—before approving a merger by its parent company.
For his part, Trump has bragged that he “brought free speech back to America,” then turned around and revoked visas from foreign students with political opinions he disagreed with.
Most strikingly, at the core of the current presidency is Trump’s lie that he really won the 2020 election. That lie motivated an attempted insurrection on Jan 6. 2021. Now every nominee for the current Trump administration must support the lie, at least in public—never acknowledging that Trump lost in 2020, but accepting only that Joe Biden was “duly sworn in.”
Based on the past actions of the president and his supporters, it’s safe to assume that AI programs will be barred from government contracts if they acknowledge that slavery is a stain on the nation’s history, that trans people exist, that Abrego Garcia was wrongfully deported, that anthropogenic climate change is real, that Biden legitimately won the 2020 election… The list could go on and on.
—Philip Athey
AI Action Plan
Top AI Action Plan takeaways:
- The AI Action Plan rolled out last week is largely seen as a major giveaway to the domestic AI industry, with supporters arguing it is necessary to defeat China, while civil society organizations complain it has given too much away to tech companies, risking greater harm to society.
- As part of the plan, Trump aims to introduce more AI into K-12 education, with the hope that it will improve outcomes and prepare children for a future in which AI technology is ubiquitous.
- Many free speech advocates warn that the executive order against “woke AI” is a way to force the current administration’s ideology into AI programs.
‘Whatever it takes to lead the world’: Trump unveils AI program: The president signed three executive orders Wednesday aimed at speeding up the environmental permitting process for AI infrastructure, reducing Biden-era restrictions on AI exports, and restricting government acquisition of AI to models that agree with the Trump administration’s views on “objective” truth. The executive orders are part of a larger AI Action Plan the White House also released on Wednesday, which calls for reducing the regulatory burden on big tech while accelerating government adoption of the technology. (National Journal)