The Insight: AI in the White House

The rest of this blog is for Sustaining Members only.

Become a Sustaining Member

Already have a sustaining membership? Sign in

After making the deal in summer of 2025, leaders at Anthropic worked to negotiate contracts with the DOD and define the terms of the technology use. After the DOD gave the AI company an ultimatum, according to CNBC (Lean Left bias), Anthropic “was booted for demanding that its AI not be used for autonomous weapons or domestic surveillance.”

In a statement on Anthropic’s website, the company wrote, “We support the use of AI for lawful foreign intelligence and counterintelligence missions. But using these systems for mass domestic surveillance is incompatible with democratic values.” It also wrote, “Frontier AI systems are simply not reliable enough to power fully autonomous weapons. We will not knowingly provide a product that puts America’s warfighters and civilians at risk.”

In response, the Assistant to the Secretary of Defense wrote on X, “The Department of War has no interest in using AI to conduct mass surveillance of Americans (which is illegal) nor do we want to use AI to develop autonomous weapons that operate without human involvement. Here's what we're asking: Allow the Pentagon to use Anthropic's model for all lawful purposes.”

A senior official at the DOD also said, “It is the military’s sole responsibility to ensure our warfighters have the tools they need to win in a crisis, without interference from corporate policies.”

On March 5, the Pentagon designated Anthropic a “supply chain risk,” and on March 6, a memo was sent to all senior leadership instructing them to remove all Anthropic products from their systems. 

In the midst of negotiations between the Pentagon and Anthropic, the rival AI company OpenAI approached the DOD for a deal using its technologies. Under the new deal struck with OpenAI, the company allowed all lawful uses of its systems. 

Sam Altman, OpenAI’s chief executive, said, “In all of our interactions, the DoW displayed a deep respect for safety and a desire to partner to achieve the best possible outcome.”

How Does the Current Administration Define a “Truthful” and “Unbiased” AI?

President Donald Trump signed an executive order in July over concerns of “woke AI” and the building of AI models with ideological biases or social agendas.

The order recognized the critical role AI will play for Americans and explained the importance of reliable outputs from these systems. 

While the order is intended to procure AI models that will not “sacrifice truthfulness and accuracy to ideological agendas,” it is only intended for models utilized by the federal government that“should be hesitant to regulate the functionality of AI models in the private marketplace.” 

According to the University of Maryland’s School of Public Policy, a bipartisan majority of 79% favor requiring any new decision making AI system with the potential to significantly impact people’s lives undergo a government-run test before being deployed. These tests would serve as a preventative measure “similar to how the government requires testing on new drugs.”

However, some have asked “Who governs government AI?” A Pew Research Study showed that only 44% of Americans trust the government to regulate AI, while 47% expressed distrust. 

President Trump, on the same day as his “woke AI” executive order, repealed some of former President Joe Biden’s orders pertaining to AI and unveiled an AI policy initiative called “Winning the Race: America’s AI Action Plan.” 

This plan instructs the National Institute of Standards and Technology (NIST) “to eliminate references to misinformation, Diversity, Equity, and Inclusion, and climate change,” adding that federal agencies may only utilize AI models that are “objective and free from top-down ideological bias.”

Tom Wheeler, writing an opinion for Brookings Institution (Center bias), said these orders “politicize AI.” Wheeler added that it is “a limitation of free speech and a violation of the First Amendment,” saying “This sounds like top-down censorship from the president to restrict anything that doesn’t align with his ideological preferences.”

“The next fight over freedom will run through AI models, according to Donald Kendal, an opinion writer for Blaze Media (Right). Kendal asserts that “constitutional limits [are] more urgent than ever,” considering “these systems will shape thought, speech, and incentives.” 

The Trump administration has said “America will not choke innovation with red tape.” Kendal agreed with White House AI adviser David Sacks who said, “We don’t like seeing blue states trying to insert their woke ideology in AI models.” 

Kendal believes the goal should be straightforward: “Build an American AI future in which freedom is embedded from the start, and constitutional guardrails shape the systems that will increasingly shape us.” 

How Much Energy Do Data Centers Use?

With the rise of AI, the need for new AI data centers is crucial. Yesterday’s data centers were simple, secure, centralized physical facilities that housed the IT infrastructure for an organization, including servers, data storage systems, and networking equipment. 

However, the data centers needed for today’s training and operations of AI technology are massive in comparison.

Deemed by some as the “new factory of the AI era,” these new facilities often boast more than 1 million square feet in building space, and several hundred acres of land; usually in rural or industrial areas where high power demands may be met. 

According to Lawrence Berkeley National Laboratory, data centers currently only account for roughly 4.4% of total US electricity consumption. However, these figures are projected to rise to between 6.7% and 12% in 2028. 

To operate, these data centers also need large amounts of water. Lawrence Berkeley estimated that in 2023, data centers consumed nearly 211 billion gallons of water, but noted that most of the water consumption was not for cooling the facility or servers, but rather for generating electricity.

These consumption demands may have direct impacts on the communities that host these new centers.

For instance, in Pennsylvania there are currently 37 active data centers, nine in the process of being built, and more than 168 in the planning stages. Some residents in PA, as well as other states are concerned with the effects this may have on local water supplies, electrical demand and higher household electric bills, and noise pollution. 

Who should regulate AI's limits?

FEDERAL GOVERNMENT: A single, national framework is the only way to ensure consistent, enforceable AI standards across the US.

STATE GOVERNMENT: States are the true laboratories of democracy, moving faster than Congress, able to respond to local needs, and have already produced landmark AI protections.

INTERNATIONAL BODIES: AI is a global technology with global consequences. Only international frameworks can prevent regulatory arbitrage, harmonize standards, and address existential risks that no single nation can contain alone.

PRIVATIZED/SELF-REGULATION : The technology industry, with its deep technical expertise and economic incentives to maintain trust, is best positioned to set and enforce AI standards.

PUBLIC INTEREST GROUPS/PUBLIC: Neither governments nor corporations can be trusted to protect the public from AI harms without independent civil society oversight. Only democratic accountability and public participation will ensure AI serves everyone.

Is it necessary to compete with China in the global AI race? 

Artificial intelligence has become central in US-China geopolitical competition. Policymakers, tech leaders and national security experts continue to frame AI development as a global race that could shape economic and military power, and political values for decades. 

Others argue that treating AI like a zero-sum game could create more risk than benefit, specifically for its environmental impact and the potential implications on the job market. How necessary is it for the US to win the global AI race?

Those who support a competitive approach argue for its strategic advantages and national security benefits, including the ability to influence economic productivity, improve cybersecurity and create advanced weapons. Several policy makers across the political spectrum have warned that if China wins the “global AI race”, its systems may reflect authoritarian policies like centralized control and censorship over democratic norms. They argue this could also give China a greater influence over global data flows and broader access to the world’s data to train AI systems. 

AI experts have also warned of China’s rapid AI development. A Fortune (Center) article said surging AI demand in the US is colliding with a fragile power grid whereas China considers that a “solved problem.”

As political leaders continue to emphasize urgency around AI development and China’s capabilities, some analysts wonder if the idea of a global “race” is overstated and could distract from other realities of technological development.

The New York Times (Lean Left) suggested AI is creating a new digital divide between countries with the ability to create cutting-edge AI systems and those that rely on importing them. In this competition, the main players are limited–the US, China and the European Union. 

A piece in Just Security (Not Rated) suggested that rather than resembling a 21st century Manhattan Project-style race, AI may come down to how countries learn to bridge the gap between invention and widespread social adoption. 

The Hill (Center) shared a recent poll showing 46% of Americans have a “somewhat negative” or “very negative” view of AI compared to 26% who had a positive view of it. As the AI race continues, leaders will continue to face rising concerns about the push to prioritize speed over safety. 

A third alternative to global competition or taking a backseat is the opportunity for international cooperation. Some advocates encourage cooperation to help mitigate unintended consequences like AI misuse or global economic disruption. Ideally, they say, the US and China would work together on safety standards and emergency communication systems to prevent negative outcomes.