America.gov - the AI replacing government

Hi, and happy Thursday.

On Tuesday, Donald Trump led a parade of tech moguls out onto the White House driveway for an impromptu press conference. 

Present were Nvidia’s Jensen Huang, Meta’s Mark Zuckerberg, Google CEO Sundar Pichai, Anthropic CEO Dario Amodei,  OpenAI President Greg Brockman and Elon Musk - among others.

It looked like the cast photo for the final season of capitalism.

(The billionaires building AI say that, once their vision is fully achieved, we’ll no longer need money).

They had gathered together on the driveway to reassure Americans, who have grown increasingly weary of AI and its risks, of the benefits of AI. 

Let’s see how they did.

The AI replacing government

Earlier that day, Trump launched America.gov - an AI chatbot intended to bring together some 29,000 federal websites under one interface.

Instead of searching for the right agency,  downloading PDFs and then completing and submitting forms that require a legal degree to understand, you simply explain to the chatbot what you need. Eventually, America.gov is intended to provide a range of government services, such as passport renewals.

I couldn’t help but notice the parallels to the predictions from  AI 2027 - an alarmist and somewhat controversial paper that painted a picture of how superhuman AI might “take over”.

AI 2027’s frightening idea was not that a robot would seize the White House. It was that the government would become dependent on a private AI company whose technology was too strategically important to slow down. (America.gov is powered by Gemini and Grok.)

In the AI 2027 scenario, tech executives advise the president. Competition with China makes restraint seem impossible. A joint government-industry oversight committee appears after the company has already become extraordinarily powerful.

It was speculative, but Tuesday produced uncomfortable visual rhymes.

And America.gov may be how the mundane version begins.

It doesn’t start with an AI abolishing government, but with it becoming the layer through which we experience government. First it answers questions. Then it completes forms. Then it recommends decisions. Eventually, removing it becomes impossible because nobody remembers how the system underneath it works.

AI is getting a rebrand

Back on the driveway, Trump pointed at individual tech CEOs and invited them to speak as reporters shouted questions.

When he reached Sundar Pichai, Trump offered one of the more unusual introductions ever received by the CEO of one of the world’s largest companies:

“This guy is a monster, and nobody knows him.”

Trump meant it approvingly. By “monster”  he meant Pichai is a formidable operator and enormously powerful, but also able to walk down the street without producing a small constitutional crisis.

“What a great life!” Trump continued. “To be a monster and not have to go through this.”

Pichai smiled and replied, “Hopefully a better tagline than I’m a monster.”

And taglines were the main business of the day.

Artificial Intelligence is now Super Intelligence

Trump signed an executive order banning the use of the term  “Artificial Intelligence” and “AI” in government communications - it must now be called “Super Intelligence,” or “SI.”

“We’re using the word super because the other word is a fake word,” Trump explained.

The rebrand reveals something real: America has an AI marketing problem.

People associate AI with lost jobs, deepfakes, surveillance and chatbots that confidently invent court cases. The industry needs them to picture medical discoveries, economic growth and someone else completing their passport application.

So the administration has responded by giving the technology a name normally reserved for the point in a science-fiction film when the computer locks the doors.

Foxes, meet henhouse

The tech executives also signed the Joint Commitment on Frontier Responsibilities. The companies will institute internal controls, independent audits and board-level oversight of frontier AI systems.

Trump praised the industry’s “tremendous self-policing.” 

There is just one problem with that sentiment.

The history of self-policing is largely a trail of disasters, but I have to admit that it’s a difficult line to tow.

Whether in food safety or high finance, we rarely discovered which safeguards were needed until a catastrophe revealed them. From the unsanitary meatpacking that birthed the FDA in 1906 to the predatory lending that triggered Dodd-Frank in 2010, self-regulation didn’t prevent disasters - it just supplied the evidence required to write the rules.

Trump suggested a committee of roughly ten people could “watch over the whole enterprise.”

Perhaps we need to put a little more thought into it than letting ten foxes self-police a henhouse.

What do you think?

Do you feel assured?

We don't either - which is why (in the systems we build) we don't rely solely on what Big Tech delivers out of the box.

We view it as our job to bridge the gap - adding the operational transparency and expert oversight that gives us confidence in standing behind the systems we build, and gives clients the trust they need.

Best,

Dino