A version of this article was originally featured in a weekend brief newsletter for only Quartz members. Quartz members have access to exclusive newsletters and more. Sign up here.
The new AI rules for Europe are here – and they’re coming to Silicon Valley.
The groundbreaking law that became law last year The deployment has begun Earlier this month, a provision banning certain “unacceptable risk” AI applications, including social scoring systems and manipulated AI technologies. Violation of the European Union’s AI Law It could result in a fine Up to 7% of global revenue or 35 million euros (nearly $37 million), whichever is greater.
The rules spark a transatlantic showdown, with the Trump administration pledging to protect American tech companies from overreach in Europe. However, the US tech industry is already preparing for compliance. Similar regulations are also gaining momentum in state capitals across the United States.
This conflict highlights the growing disparity in approaches to AI governance. The EU has chosen comprehensive regulations aimed at ensuring that AI development is consistent with European values and rights, but the US has not passed anything at the federal level. Former President Joe Biden only Issuing an executive order We have outlined voluntary AI safety guidelines. Even that minimal effort was rewinded by President Donald Trump, who advocates minimal surveillance to promote innovation.
“These are American companies, whether you like it or not.” Trump said Last month, crowds at the World Economic Forum in Davos characterized the EU’s regulatory approach as a “form of taxation.” His administration is increasingly confronting with the Vice President. JD Vance warning at Paris AI The summit is where foreign governments “cannot and will not accept” that the US is “tightening screws on high-tech American companies.”
Big Technology leaders reflect these concerns. Meta CEO Mark Zuckerberg criticized what he called “Institutionalized Censorship” in Europe; Meta’s policy chief Joel Kaplan The company has shown that it seeks Trump administration Intervention when people feel unfairly targeted by EU enforcement.
European officials maintain that regulatory frameworks are necessary and fair. “When we’re doing business in other countries, we have to respect their rules,” said EU committee member Henna Wilkunen. Emphasise that rules apply Similarly, for businesses in the US, Europe and China. However, the pressure seems to have some effect. Virkkunen also told Reuters The Commission shall examine the rules to eliminate duplicate regulations.
“We will reduce the deficit and administrative burden from our industry,” she said.
Civil society organizations are urging the EU to stand firmly. Alliance of NGOs warned recently The European Commission president has been referring to the EU’s broader technical regulatory framework, “to undermine DSA and DMA enforcement by being bullied by things like (Elon) Musk and Trump.”
Politicians trade barbs in the Atlantic, but reality looks quite different on earth. There, companies are already thinking about ways to follow AI laws.
“It doesn’t matter where the company is. It’s important to know where the products, services and consumers are,” said Reyidgani, a professor at Heinz College at Carnegie Mellon University. “If they touched EU people or residents in any way, that applies to them. Companies know that very well.”
The first major deadline to hit on February 2nd Staff AI literacy Working with AI products. Ghani said this initial requirement could not dramatically change the way companies build their products – partly due to a lack of established training programs, but companies are already preparing for the more substantial requirements that will come in 2027.
“They do the exact same kind of things they did for the GDPR,” Gani said. “They adopt the inventory of AI systems they have, classify each piece by risk level and know how to monitor and ensure compliance.”
Meanwhile, regardless of the Trump administration’s stance, US states are not waiting for federal actions.
“The nation is considering precedents,” Ghani said. “That was the first thing, so they’re looking at EU AI law, but they’re looking at the AI risk management framework and Biden’s executive orders.”
More 12 states – including California, Texas, Virginia and New York – is considering or implementing legislation that focuses on important aspects of EU AI law, particularly “algorithmic discrimination” in automated decision-making systems. Colorado became the first state to pass the comprehensive AI law last year, with others posing for bills requiring risk management plans and impact assessments for AI systems used in “high-risk” contexts such as employment, education and financial services.
Full EU regulations will not begin until 2027, but this patchwork of state law could reconstruct the American AI landscape much faster. And Silicon Valley may have a hard time fighting multiple battles in the backyard than taking on Brussels.
“The states are trying to figure out what to do. Things are horizontal, things are vertical and they are able to pass,” Gani said. “Especially now, they’re realising things don’t move at the federal level.”
