VANCOUVER, BRITISH COLUMBIA - 

This article has been adapted from the original version, which can be found on Gary Marcus' Substack. 

The European Union must cut a deal soon, because its pioneering, world-leading AI Act is on the ropes.

In the worst case, five years of negotiation could lead to nothing, leaving EU residents as well as the rest of the world more or less entirely on the hook for any negative externalities that arise from generative artificial intelligence (GenAI) - from misinformation to cybercrime to new bioweapons to bias. The ‘less’ in this ‘more or less’ implies that some existing regulations cover some tiny fraction of the conceivable negative externalities, but since virtually all existing laws were drafted before artificial intelligence (AI) really came onto the scene, very few truly envision how AI changes things. Runaway AI - if that is in fact a thing - would also not be covered.

Big Tech wants the world to revolve around its new plaything, GenAI - whose combined total revenue so far is just a few billion dollars, a tiny fraction of the world economy - and to exempt it from any regulation.

This should not be allowed to stand.

Several people in the industry - including Canadian computer scientist Yoshua Bengio, ‘godfather of AI’ Geoff Hinton, British professor and computer scientist Stuart Russell, Dutch politician and international policy expert Marietje Schaake, and myself - recently signed a pair of open letters urging policymakers to seek a compromise. One letter was addressed to the German government, while another was written more broadly to the world at large.

The one to the German government stressed:

“[I]t is vital that risks are addressed at the foundation mod