HATIM KAGHAT / BELGA / AFP)
WASHINGTON: US President Donald Trump’s administration is putting pressure on the Europe Union (EU) to ditch a rulebook that would compel developers of advanced artificial intelligence (AI) to follow stricter standards of transparency, risk-mitigation and copyright rules.
The US government’s Mission to the EU reached out to the European Commission to push back against the AI code of practice in the last few weeks.
The letter argues against the adoption of the code in its current form, and also went out to several European governments, people familiar with the matter said.
In response to Bloomberg questions, commission spokesman Thomas Regnier confirmed the reception of the letter.
While the code, which is still being finalised, is voluntary, it’s meant to give tech companies a framework for staying in line with the EU’s Artificial Intelligence Act.
Running afoul of the AI Act carries fines of as much as 7% of a company’s annual sales.
Fines for the developers of advanced AI models can reach 3%. And not following the code could mean greater scrutiny from regulators.
Critics have said the guidelines go beyond the bounds of the AI law, and create new, onerous regulations.
Trump and the EU are increasingly clashing over the latter’s role as the world’s chief digital rule maker, and the US president has lambasted the bloc’s tech regulations and fines as unfairly targeting American companies.
House Judiciary Committee chairman Jim Jordan sent a volley of letters to EU leaders asking them to respond to his concerns that their policies were restricting Americans’ free speech rights.
Trump has said the bloc’s tech regulations are “a form of taxation”, in comments at the World Economic Forum in Davos in January.
The US letter called the code excessively burdensome, and detailed its concerns about the code of practice, and the AI Act at large, one of the people said, asking not to be identified because the letter isn’t public.
The letter went as far as requesting that the EU put the entire AI Act’s multi-phase implementation process, whose centrepiece is the code of practice hemming in powerful AI models, on hold unless the issues raised by the US are addressed, the person said. — Bloomberg
