The new Sanders-Casar Ban Artificial Superintelligence Act – and why I oppose it

Substack · Watch/Read · (gathered 2026-09-04)

• Gary Marcus opposes the proposed Sanders-Casar “Ban Artificial Superintelligence Act,” arguing that while a temporary pause on advanced AI development and an AI agency might be necessary, a permanent, unilateral ban on all research into superhuman AI is too broad and detrimental.
• Marcus contends the proposed ban is naive about benchmarking, risks leaving the US behind, and focuses excessively on hypothetical future risks (like AI disempowering humanity) at the expense of addressing current, real-world AI dangers, as highlighted by Dr. Angela Rasmussen.
• He suggests an alternative framework, inspired by Anthony Aguirre, where narrowly-scoped AI development can continue, but systems of greater intelligence, generality, and autonomy are permitted only if developers can demonstrate “high assurance to a capable independent authority that sufficient alignment, control, and oversight exist.”
• The current proposal is “not the way” to achieve a good AI future, as it’s overbearing and could stifle beneficial research; instead, a nuanced approach that incentivizes responsible development through high standards is preferred.