You can’t address AI’s risks by banning ASML’s exports

In a recent opinion piece in Dutch daily De Volkskrant, Kritika Maheshwari and Otto Barten argue for far broader export restrictions on ASML’s lithography systems (link in Dutch). Under their proposal, only organizations that “respect the rule of law and develop AI in a peaceful, democratically controlled, sustainable and safe manner” should be allowed to buy the company’s most advanced scanners. Their primary target is Elon Musk and his planned chip venture, Terafab.

Maheshwari’s and Barten’s concerns are justified. AI is advancing faster than governments can establish meaningful democratic oversight – assuming they’re even willing to try. As things stand, society isn’t adequately preparing for the technology’s risks (nor for its enormous potential).