Gary Marcus opposes Sanders-Casar Ban Artificial Superintelligence Act as too broad

GaryMarcus · x · 2026-09-04

Gary Marcus published a free Substack post explaining why he opposes the Sanders-Casar Ban Artificial Superintelligence Act, which would pause all US AI development, ban systems matching or exceeding human cognitive performance, and impose up to 20 years in prison for violations.

Marcus sees merit in parts of the proposal but argues it clearly goes too far
He supports a temporary pause—possibly lasting years—and has previously testified for an AI agency, but a permanent unilateral ban on superhuman AI research would guarantee the US falls behind
The bill is "naive about the complexities in benchmarking" and leaves room for authoritarians to game the rules
It overweights hypothetical future risks (AI "planning to disempower humanity") at the expense of current risks
In follow-up posts he adds further reasons and positive suggestions inspired by @AnthonyNAguirre

Original post →

Gary Marcus opposes Sanders-Casar Ban Artificial Superintelligence Act as too broad

More from Safety