“Whoever has the largest AI ecosystem will set the global standards and reap broad economic and security benefits. Under President Trump, our Nation will win, ushering in a new Golden Age of innovation, human flourishing, and technological achievement for the American people.”
On July 23rd, 2025 Trump released this statement on a government website with an introduction to the ‘AI Action Plan,’ an outline of policy directives that will guide the U.S. government’s use of artificial intelligence. Secluded in this triumphant language is a mentality that reverberates through time: mercantilism. A dominating economic theory of the 16th through 18th centuries, mercantilists believe that the wealth in the world is finite; every iota of wealth our country gleams is wealth now unavailable to other countries. The wealthiest country is thus best suited for global success, so countries should minimize imports and maximize exports. By analogy, the U.S. government seeks to maximize its influence and control over the AI industry worldwide while minimizing the influence of close competitors like China, through policy.
This couplet of sentences also microcosmically encapsulates the psychological pathos of the MAGA movement. It calls for a nostalgic and triumphant return to a “Golden Age” (hence Making America Great Again). Rather than the Republican party, the U.S. government as a whole, a set of ideals, or the people of America themselves, the sentence frames President Trump as the illustrious benefactor responsible for America’s imminent prosperity, centering him as the figurehead of the movement.
The AI Action Plan aims to fulfill this grandiose mission by outlining the three pillars necessary for success.
The first pillar is “Accelerating Innovation.” Among the several initiatives outlined to accomplish this mission, Trump has intended to mobilize several agencies to remove bureaucratic barriers, encourage the availability of “Open Source” models and datasets that anyone can download, provide economic incentives to benefit AI startups, and integrate AI systems into government and education.
The second pillar is to “Build AI Infrastructure.” This pillar entails providing economic incentives to build more AI data centers, and supporting the manufacturing of such centers by optimizing energy use in order to support the heavy energy demands of AI, and improve semiconductor manufacturing. Semiconductors are the chips used in AI technology. Job training will also be provided for this maintenance of infrastructure and for cybersecurity.
“Leading in International AI Diplomacy and Security” constitutes the intent of the Plan’s third pillar. It entails selling America’s AI technology to its allies and avoiding sharing it with direct competitors, preventing national security risks, and exploring its potential uses in biology. This section directly names China as a foreign adversary that America is competing with in AI development and preventing Chinese influence in international governance bodies concerning AI.
Appointed government officials and an outpouring of money could be influencing this decidedly pro-AI agenda.
David Sacks acts as the Trump Administration’s AI and cryptocurrency czar, an expert that will influence federal public policy, given his specialties. However, news outlets have noted that Sacks has invested in 449 companies involved in Artificial Intelligence. Because he is a government official, he could be aided directly or indirectly by the financial outcomes of the companies he personally invests in. Sacks has also attempted to use his government position to popularize his podcast, All-in. Most notably, he planned for ‘All-in’ to hold the US’s July 2025 AI summit, requiring sponsors to pay $1 million for entry to a private reception between Donald Trump and leaders in AI before the idea was struck down by White House Chief of Staff Susie Wiles.
AI companies have also been lobbying the government to form government contracts and prioritize less regulatory policy to benefit their businesses. The top 5 companies affiliated with AI involved in lobbying, Meta, Google, Oracle, Microsoft, and Palantir, donated over $70 million in 2025 alone. Leading The Future is an AI super PAC valuated at $100 million dollars that is currently funding candidates that will likely prioritize their ideal legislation in this upcoming midterm election.
Trump’s “Big, Beautiful Bill,” passed on July 4th, 2025, enshrines the directives outlined in the AI Action Plan into U.S. governance through policy. Approved in the House of Representatives, the bill initially included a proposition that would prohibit state-level AI regulatory for ten years, in order to accelerate innovation by streamlining AI companies’ pathway to success without concern for state-by-state inhibitions.
This provision of the bill was struck down in a 99-1 vote by the Senate. What changed? The opposition argued that it prioritized company profits over the American people and their agency in being able to control how AI could and does affect their lives. As Democratic Senator Ed Markey of Massachusetts stated, “This 99-1 vote sent a clear message that Congress will not sell out our kids and local communities in order to pad the pockets of Big Tech billionaires.”
Other provisions in the bill may contradict that messaging. Because of the efforts of these companies for involvement in government, they have enjoyed several benefits. For example, an Institute of Taxation and Economic Policy analysis reported that Palantir would have paid $330 million in taxes in 2025 if not for the Big Beautiful Bill’s tax cut pathways, since it would have paid 21% of its profits as the federal income tax rate as a company that profited over $1.5 billion dollars.
The bill also states that, in alignment with “Leading in International AI Diplomacy and Security,” the U.S. government will provide federal grants and funding to companies developing AI infrastructure, contingent upon their use having no involvement with “prohibited foreign entities,” namely China. That involvement includes licensing, supply chains, and other methods that U.S. technology could be transferred to foreign countries. That means that companies are strongly incentivized to create teams in order to track and document their entire supply chain and involvement, to ensure that they can continue using their expenditures as federal tax credits.
These concerns over national security are not unfounded. Anthropic discovered in November of 2025 that China modified its Claud AI tool to launch a cyberattack, which could have resulted in China taking data from U.S. businesses and consumers.
When Deepseek released its R1 generative AI reasoning model in January of 2025, it shocked pop culture and the stock market, as it seemed to rival ChatGPT’s capabilities despite using a fraction of the power. While the actual merits of Deepseek over ChatGPT are debated, the dominating narrative became that the U.S. had seriously underestimated China and that China’s AI advancement was legitimately competitive with the U.S. ‘s. The day of its release, the stock markets plummeted, with Nvidia losing approximately 600 billion, 17% of its total market share in a day. Though these kneejerk reactions have long since stabilized, the memory plants a firm urgency in American AI development.
However, the adherence to these strict rules in the name of preventing sharing of AI technology with foreign adversaries seems situational. Trump recently allowed Jensen Huang, the CEO of Nvidia – America’s biggest AI Chip company– to sell the advanced new H200 chip to China. The contingencies are that the U.S. is to receive 25% of the profits from the chips sold to China, and that Nvidia’s most powerful computers the Blackwell and Rubin GPUs remain off limits. Regardless, it reveals a contrast in some actions to the language and policy directives being outwardly popularized.
Much of the collaboration between AI companies and the U.S. government comes in the form of government contracts. Collaboration with the government has continued to boost investing in AI in a positive feedback loop.
With Nvidia alone being valued at over $4.6 trillion dollars, the most of any company, AI is currently driving the economy. AI companies are currently participating in circular investing, in which Nvidia, Microsoft, OpenAI, and other AI tech companies invest in each other and buy each other’s exports, creating an inflated view of productivity. Many people are calling the speculative investment into AI a bubble, including the CEO of OpenAI. Sam Altman himself said in an August 2025 interview from The Verge, “Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes.” Likewise, Sundar Pichai, CEO of Alphabet, Google’s parent firm, has said that the AI investment boom has “elements of irrationality,” as valuations and investments keep pouring in.
When the leaders of the “AI revolution” admit there are concerns with over-evaluations of their own companies and any government intervention seeks to only increase AI’s involvement in the economy, the reaction and cause for concern is understandable. This, coupled with the fact that these major AI companies are unprofitable in comparison with their investments, has caused many, including Sam Altman himself in his interview to draw parallels between the current economic landscape and the “dot com” bubble of the late 1990s, for which investors poured money into unprofitable business models that emerged with the introduction of the internet. During the dot com crash, the NASDAQ index lost approximately 77% of its value as internet startups with unprofitable business models that had received heavy investments went bankrupt or lost significant funds. Concerningly, Chief Economist of the private equity Apollo Global Management Torsten Stok found that the AI bubble is even larger than the dot com bubble was in its over-evaluation of top companies for the particular industry. The dot com and AI bubbles are not necessarily an exactly analogous comparison because of some differences.
The justification for this outpour of investing in AI is to help current AI technology reach artificial general intelligence (AGI) that can independently think similar to or better than a human. AI infrastructure would hypothetically then become extremely profitable and beneficial to humanity. Some contend that it’s a practical impossibility, others believe that even if possible, it wouldn’t create the utopian reality promised and that such a strong economic dependence on AI is counterproductive.
The existentialist pessimism over the AI economic situation is only a facet in which younger generations have a complicated relationship with AI use in different companies. A Stanford survey found that the United States was second only to Belgium in terms of being the country least excited about AI.
HAL 9000 betrays the crewmates in the Space Odyssey in its programming in Stanley Kubrick’s classic film from 1968, 2001: A Space Odyssey. Humans are imprisoned in a virtual reality in the 1999 film, The Matrix. In the movie The Terminator (1984), the terminator is a self-aware AI bot intent on destruction. A distrust for AI has been deeply ingrained into American pop culture for decades. What happens when that distrust contends with the reality that the U.S. government, companies, and other powerful institutions are persistent in positioning it as the key to the future?
The language learning app Duolingo serves as a case study for how these anxieties intersect with pop culture and consumerism. When Luis Von Ahn, CEO of the Duolingo announced in a Linkedin post that his company would transition to an “AI-first” approach, shifting to using AI to generate the practice lines of text and limiting future hiring of human workers. This announcement incited a wave of backlash from people who believed the shift would compromise the value of the product because they believed the “AI-first” approach replaced workers who wrote educational content, creating a less human and less effective product. Duolingo’s green bird mascot, Duo, had been the centerpiece of their marketing campaign and frequently inserted itself into popular internet trends. Immediately after the Llinkedin post, all the comments on Instagram related to this topic consisted of reels and posted comment sections flooded with a GIF gift of a red sign crossing out the word “AI.”
Simultaneously, generative AI art and images haves also drawn ire. Critics exclaim that AI art does not credit the images it took inspiration from and doesn’t have the consent of other artists to use their art, does not involve the intensive effort or decision making that makes art distinctly human and valuable. Furthermore, it oftentimes does not self-disclose that it was generated with AI. A September 2025 Pew Research Center survey assessed the public’s reactions to AI, finding that half of the surveyors thought AI would worsen people’s abilities to think creatively.
As AI images and videos prioritize realism, distinguishing hyper-realistic AI videos and images from reality becomes increasingly difficult. As per Pew’s survey, for at least 76% of Americans, it is imperative to be able to distinguish AI from man-made creations, yet 53% show little to no confidence in their ability to do so. Just recently on August 26th, 2025, Google’s Nano Banana Pro launched. The tool is one of several services able to create hyperrealistic images. This hyperrealism raises concerns about how AI could be used to spread misinformation. Some image generators embed watermarks in the images they generate, but a “watermark remover” that is accessible from a single google search can remove the trace. Stricter regulation on disclosure if content is AI generated could be a potential solution to this problem.
Besides misinformation, AI bots have also perpetuated racial biases. For example, Nano Banana Pro disproportionately represents white women providing aid to black children when Banana pro is prompted to depict humanitarian aid in Africa, an example of the “white savior” bias in these algorithms. AI models can be tweaked when this behavior is found, but it is often more covert, and issues are remedied only once they are found. Models pick up these biases from the data they analyze, data labeling, and other parts of the AI creation process. If we continue to use AI uncritically without thinking of how bias can and does manifest itself in AI generated content, that prejudice will be perpetuated.
The U.S. has also ended its contract with Anthropic for the U.S. government and military. Anthropic declared it would not remove safety restrictions against using its AI model, Claude, for mass surveillance and with fully autonomous weapons, breaking off its deal with the U.S. Government over the issue. AI has the potential to “hallucinate” wrong information which, in a worst case scenario, could lead AI to target the wrong people, and facilitate mass domestic surveillance.
President Trump commented on the fallout of the deal with the military in a Truth Social post, calling Anthropic employees behind the decision “Left-wing nut jobs,” trying to “STRONG ARM the Department of War” to “obey their Terms of Service instead of the Constitution… Their selfishness is putting AMERICAN LIVES at risk, our Troops in danger, and our National Security in JEOPARDY,” and declaring that the U.S. military would completely phase out Anthropic technology within 6 months.
Anthropic CEO Dario Amodei commented on the fallout, saying that AI can be used to “assemble scattered, individually innocuous data into a comprehensive picture of any person’s life — automatically and at massive scale,” which is “incompatible with democratic values.” By drawing “red lines” against AI being used for dubious military uses, he patriotically stood up for American democratic values.
Anthropic’s decision has changed people’s perceptions of it as a the most moral AI company. It’s already known that Claude was used for the strikes in Venezuela in order to capture Nicholas Maduro which is reportedly where tensions began since this act conflicted with Anthropic’s ideals of use of AI for violence. An Anthropic spokesperson said that “virtually no progress on preventing Claude’s use for mass surveillance of Americans or in fully autonomous weapons.” The U.S. military reportedly used Claud in the recent strikes in Iran and Venezuela.
The Department of War has since labeled Anthropic a “supply chain risk” on March 4th 2026, a designation that has previously only been applied to technology from foreign countries, which could interfere with it contracting with other parts of the U.S. government in the future. Anthropic has since sued the Department of Defense in two courts, accusing it of violating due process and free speech by targeting the company because of ideological differences, though it is still open to renegotiation.
Anthropic’s decision-making juxtaposes with those of other AI companies who have contracted with the U.S. government more liberally.
In the same vein as nuclear weapons, just because a government can develop a lethal technology that could give it a military advantage, doesn’t necessarily mean it should.
Chief Pentagon technology officer Emil Micheal told CBS that “At some level, you have to trust the military to do the right thing” with its use of AI. Many citizens, however, feel that Anthropic made the right decision by declaring that it would not budge on artificial intelligence used for military violence.
Meanwhile, OpenAI has stepped in to effectively replace Anthropic since the fallout. Already, Palantir has partnered with the U.S., Ukraine, the United Kingdom, and NATO for them to use their data analytic tools.
A lot of interpretation of AI policy and AI in general comes down to framing. As with any issue, read about this issue from a variety of sources to come to conclusions and advocate for action yourself, and be cognizant of the moneyed institutions seeking to socialize you to accept their status quo. Much of the news that you will hear, especially from the U.S. government or the press releases from the companies themselves will frame AI positively. These are powerful interest parties that benefit from normalizing AI in our everyday consciousness. Tax cuts give companies the resources for innovation, or they are the result of corporate lobbying and take away money that could be used to help everyday Americans or reduce our $38.8 trillion in national debt. Depending on your viewpoint, contracts between the U.S. government and the military promote national security and innovation, or they are reckless. Depending on your viewpoint, Anthropic is acting against national security by breaking with President Trump, or Anthropic is a patriotic company that is standing for what is right.
The reality is that we are never going back to a world without artificial intelligence, so we should ensure our government and companies use it as ethically as possible for the common good.
The reality is that we are never going back to a world without artificial intelligence, so we should ensure our government and companies use it as ethically as possible for the common good.