Trump’s AI renaming order tests the tech industry’s loyalty

A name with seven decades behind it

“Artificial intelligence” dates to a 1956 gathering at Dartmouth College, where computer scientists discussed whether machines might one day think like people. John McCarthy proposed the name. It stuck: a Google Scholar search returns about 7,420,000 results.

Trump had already signaled his dislike of the phrase in a speech at the United Nations on September 23. At the White House lunch this week, he made the preference official. The executives who heard it included Jeff Bezos, Jensen Huang, Dario Amodei, Mark Zuckerberg, Sundar Pichai, Greg Brockman, Alex Karp and Elon Musk. None appears to have objected.

That silence let Trump say the change had the backing of the industry’s leading figures. When a journalist later asked him about AI agents, he corrected her and said the world’s biggest, smartest and greatest people had changed the name.

The public record offers little evidence that those leaders wanted the change before the lunch. A search using systems such as ChatGPT and Gemini turned up no instance of any of them advocating that “artificial intelligence” be abandoned. Satya Nadella once said he preferred “other intelligence,” but continued to use the familiar term. Zuckerberg has recently created a superintelligence lab; Meta still sells Meta AI to billions of users and calls its products AI glasses.

The author asked Google, Microsoft, Anthropic, Meta, Amazon, OpenAI and xAI whether they would change references to AI in product names, research papers and communications. OpenAI and xAI were also asked whether they would change their company names. None had responded by publication.

Jacob Weisberg, author of the recent book Profiles in Cowardice: A Study of Collaboration in the Trump Era, called the episode a classic Trump loyalty test: a demand that influential people abase themselves to win his favor and avoid his anger. In Weisberg’s view, executives could have resisted together. If they had all accepted self-regulation while rejecting the name change, Trump would have had to decide what to do next.

The word is not the problem

The proposed replacement also collapses distinctions that matter. Philosopher Nick Bostrom, who popularized “superintelligence” in a book of the same name, told the author by email that the word describes systems that greatly surpass the best human minds.

Artificial intelligence covers a much broader range: spam filters, spell-checkers, simple navigation systems, early large language models and 1970s chess programs that could not yet beat top players. Calling all of them superintelligence would be strange.

Trump has argued that “artificial” makes intelligence sound fake. But the word was used to distinguish machine intelligence from human intelligence, not to claim that machine outputs are unreal. The practical concern runs in the other direction: AI systems can be convincing enough that people may not know whether they are interacting with a model or another person. Clear labeling, including watermarks, should be part of AI regulation; the White House does not appear inclined to pursue that.

I think the more revealing question is not whether companies will rename their products. They may do the minimum needed to avoid Trump’s displeasure and wait for a future administration to reverse the order. The unanswered question is what their promise of self-regulation is worth when the people making it have just shown how quickly they will yield on something central to their work and identity.

Voluntary standards require companies to hold their ground when safety conflicts with commercial and competitive interests. The executives left the lunch with a promise of self-regulation in place of stricter oversight, but their silence has made that promise harder to trust.