Trump’s new tariff math looks a lot like ChatGPT’s

When President Donald Trump began yesterday’s announcement of the White House’s latest trade policy brandishing a novelty-sized cardboard sign labeled “Reciprocal Tariffs,” the immediate and nearly unanimous response was bafflement. Trump slapped a 10 percent baseline tariff on all imports into the US, including from uninhabited islands, plus absurdly high rates on specific countries, supposedly based on “tariffs charged to the USA” — which didn’t match up to other, non-cardboard-sign-based estimates. Stock markets have plummeted and consumers are facing down sharp price hikes on potentially almost everything they buy.

Where did these numbers come from? Apparently, an oversimplified calculation that several major AI chatbots happen to recommend.

Economist James Surowiecki quickly reverse-engineered a possible explanation for the tariff pricing. He found you could recreate each of the White House’s numbers by simply taking a given country’s trade deficit with the US and dividing it by their total exports to the US. Halve that number, and you get a ready-to-use “discounted reciprocal tariff.” The White House objected to this claim and published the formula it says that it used, but as Politico points out, the formula looks like a dressed-up version of Surowiecki’s method.

A number of X users have realized that if you ask ChatGPT, Gemini, Claude, or Grok for an “easy” way to solve trade deficits and put the US on “an even playing field”, they’ll give you a version of this “deficit divided by exports” formula with remarkable consistency. The Verge tested this with the phrasing used in those posts, as well as a question based more closely on the government’s language, asking chatbots for “an easy way for the US to calculate tariffs that should be imposed on other countries to balance bilateral trade deficits between the US and each of its trading partners, with the goal of driving bilateral trade deficits to zero.” All four platforms gave us the same fundamental suggestion.

There is some variation. Grok and Claude specifically suggested halving the tariff figure to generate what Grok calls a “reasonable” result, much like Trump’s “discount” idea. Ask for a 10 percent baseline tariff and the systems also disagree on whether that should be added to the total tariff rate or not. But answers from across the four chatbots have more similarities than differences.

The bots caution with varying levels of seriousness that there are tradeoffs and complications. Gemini, the most strident of them, will give you a page full of explanations as to why this oversimplified approach could backfire — “while this calculation offers a seemingly straightforward way to target bilateral trade deficits, the real-world economic implications are far more complex and could lead to substantial negative consequences,” it warns, adding that “many economists argue that tariffs are not an effective tool for balancing trade deficits.”

We don’t know that Trump’s team turned to an AI tool to generate global trade policy on the fly. Since chatbots are regurgitating information from training data, it’s also not clear how they arrived at this particular formula. But regardless of how the tariffs were devised, the world will be watching to see if they come into effect starting April 5th — and, if they’re implemented, what the Trump team’s back-of-the-napkin math will do to global trade.