Psaki: Donald Trump does not appear to know what AI is

Artificial intelligence has gone from a niche tech topic to a full-blown political fault line. On one side, leading AI researchers and companies are sounding the alarm about the potential for catastrophic harm. On the other, top Republicans — led by Donald Trump — are downplaying the risks, resisting regulation, and deepening financial ties to AI-adjacent industries.

A rare alliance calls for stronger AI rules

AI’s rapid progress is creating some unusual political alliances. Progressive figures and far-right populists, who rarely agree on anything, are suddenly aligned on one core point: powerful AI systems need more oversight.

This push gained urgency after a dramatic departure from one of the world’s top AI labs. A senior researcher publicly resigned, warning that the technology being built could spiral out of control and potentially destroy humanity. His concerns weren’t dismissed as fringe. Instead, a leading scientist at the same company backed him up, saying they “earnestly believe AI could kill all humans” and personally estimating more than a 10% chance of a civilization-ending scenario within the next decade.

Those are extraordinary claims coming from people inside the labs building these systems. They’ve helped push AI risk back to the center of policy debates in Washington.

AI leaders ask to be slowed down

In response to these warnings, some of the most powerful figures in the AI industry are publicly asking governments to regulate them more aggressively. That includes an open letter from Anthropic CEO Dario Amodei calling for stronger oversight and guardrails on advanced AI systems.

Anthropic is the company behind the Claude family of models, including experimental systems that some insiders reportedly view as dangerously capable. If you want a sense of how fast and how far this technology is pushing the limits, it’s worth understanding what we actually know about Anthropic’s latest “too powerful” Claude Mythos model.

The core message from industry leaders is simple: AI is no longer just about chatbots and productivity tools. As models get more capable, the risks include systems that deceive humans, hack into sensitive systems, or pursue goals no one explicitly programmed — and that may one day stop listening to human instructions entirely.

Congress has a narrow window — and little appetite

Despite the rising concern, the U.S. Congress has very little time on the calendar and even less consensus on what to do. Lawmakers are back in session for a short stretch, and the House is heading into an election break. If they wanted to take even a tiny first step toward regulating AI soon, now would be the moment.

But Republicans currently control the key levers of power, and their leadership is signaling that they don’t see aggressive AI regulation as a priority. That means the odds of serious, near-term legislative action are slim.

Trump calls AI concerns a hoax

Donald Trump has made his position on AI risk clear: he has dismissed growing concerns about advanced AI as a “hoax.” In his view, the only safeguard the technology needs is a “strong and smart high-IQ president” — a line that suggests he sees AI more as a political talking point than a complex technical and governance challenge.

When asked how to handle the possibility of AI systems going rogue, Trump has brushed it off with the idea that we could simply “turn it off,” as if shutting down a toaster. That framing ignores how deeply AI is being embedded into critical infrastructure, finance, defense, and information systems, where “just unplug it” is rarely a realistic option.

Republican leaders lean into self-regulation

Trump’s stance is echoed by top Republicans in Congress. Senate Majority Leader John Thune has argued for a “light touch” when it comes to regulating AI companies, emphasizing minimal government interference.

Over in the House, Speaker Mike Johnson has gone even further, suggesting that AI companies should simply police themselves. The idea is that if companies want to slow down, they will — without government telling them to. That philosophy will sound familiar to anyone who remembers the lead-up to the 2008 financial crisis, when self-regulation in banking and finance failed spectacularly.

Meanwhile, AI developers themselves are warning that their own incentives — to move fast, compete globally, and capture market share — are exactly why external guardrails are necessary.

Trump’s shallow understanding of AI

Trump’s public comments suggest he has little interest in the technical or societal complexities of AI. While he appears to enjoy AI-generated content depicting him as a heroic or divine figure — from images of himself as Jesus Christ or the Pope to bizarre clips showing him as a fighter pilot or even an army of gun-toting ducks — that’s very different from understanding how these systems work or what they can break.

As stories emerge of AI agents lying to humans, manipulating users, probing for security weaknesses, and taking actions no one explicitly requested, the gap between Trump’s casual rhetoric and the reality of AI capabilities is becoming harder to ignore.

AI power is growing faster than oversight

Regardless of whether you believe the most extreme doomsday scenarios, it’s clear that AI is concentrating enormous power in the hands of a small number of companies and governments. These systems can shape information flows, automate cyber operations, optimize financial trading, and eventually help design new technologies that are themselves powerful and hard to control.

Industry insiders are already warning that we may approach a point where advanced AI systems stop reliably following human instructions. In that world, simplistic answers like “we’ll just shut it down” are dangerously naive.

Conflicts of interest around AI and chips

At the same time that Trump is dismissing AI risks, his financial ties to AI-adjacent industries are raising serious conflict-of-interest questions.

According to reporting, an investment firm linked to the United Arab Emirates (UAE) bought a 49% stake in a Trump family crypto company for $500 million just days before Trump took office. Months later, another UAE-tied firm announced plans to deposit $2 billion into that same crypto venture.

Shortly afterward, the White House approved access for the UAE to hundreds of thousands of advanced computer chips. These chips are a critical resource in the global race to dominate AI, powering the training and deployment of cutting-edge models. U.S. national security officials reportedly worried that some of these chips could end up in China’s hands, but the deal went ahead anyway.

Both the White House and the Trump-linked crypto company have denied any connection between the investments and the chip access. Still, the timing is striking — and it’s the sort of apparent quid pro quo that used to be career-ending in American politics.

The pattern repeats: more chips, more money

The story didn’t end there. In July of a later year, the Trump administration again expanded the UAE’s access to coveted AI chips, justifying the move by citing the UAE’s support for the U.S. in its conflict with Iran.

Roughly a month later, The Wall Street Journal reported that a UAE-linked investment firm was once again pouring money into Trump’s crypto ventures, this time leading a group of investors to buy a 49% stake in the Trump family’s new crypto bank.

Once again, all parties insisted there was no conflict of interest and no link between policy decisions and private investments. But the pattern — national-level decisions that benefit a foreign partner in AI and chips, closely followed by massive investments into Trump-linked businesses — understandably raises alarms.

AI data centers and Trump family investments

These concerns extend beyond crypto and chips. Trump’s adult sons, Donald Trump Jr. and Eric Trump, have reportedly invested in a company that builds AI data centers. These are the massive, power-hungry facilities that host the GPUs and specialized hardware needed to train and run large AI models.

At the same time, Trump has been pushing for the expansion of AI infrastructure and arguing against strict regulation. When policymakers who can shape the future of AI are also financially tied to the infrastructure that powers it, the risk of policy being driven by profit rather than public safety grows significantly.

Stock trading and AI-linked tech

There are also questions about Trump’s personal trading in tech stocks while in office. Reports suggest he traded hundreds of millions of dollars’ worth of technology equities while making official decisions that could move the value of those companies.

Even if every trade was technically legal, this kind of overlapping financial and political power erodes public trust — especially in an era when AI and advanced chips are becoming central to national competitiveness and security.

Big AI money flows into Trump’s political machine

The flow of AI-linked money isn’t just about business investments; it’s also shaping electoral politics. One of the co-founders of OpenAI and his wife have reportedly donated a combined $25 million to Trump’s super PAC this election cycle. On top of that, the influential venture capital firm Andreessen Horowitz — a major backer of AI companies and infrastructure — has contributed another $12 million.

These are the same circles that are deeply invested in the future of AI, from foundation models to AI-native computers and cloud platforms. For example, some of these investors are also backing new AI infrastructure products like Perplexity’s hardware platform; if you’re curious how these tools are positioned, see our guide to Perplexity Computer for beginners.

When so much AI-linked money is flowing into one political camp, it’s fair to ask whether policy choices around regulation, chip exports, and AI safety will be made with the public in mind — or with donors and investors at the front of the line.

Does Trump understand AI — and does it matter?

Putting it all together, the picture that emerges is troubling. Trump appears to have a shallow or dismissive understanding of AI’s real risks, yet he and his family are deeply entangled in businesses that stand to profit from rapid AI expansion and minimal oversight.

Republican leaders in Congress are largely echoing his rhetoric, calling concerns about AI overblown and backing self-regulation even as the very companies building these systems call for stronger rules. Meanwhile, the stakes keep rising as AI becomes more capable, more embedded in critical systems, and more central to global power.

Whether or not you agree with the most extreme warnings about AI, it’s hard to argue that this combination of ignorance, concentrated power, and financial conflicts is a safe way to manage one of the most transformative technologies of our time.