Is anyone prohibiting ASI?
Treaties, laws, institutes and pledges on ASI, listed by jurisdiction: what is in force, what is voluntary, and whether any of it prohibits building superintelligence.
The most common argument against slowing down is: “It doesn't matter what we do, China will race ahead.” The record complicates that story. China issued binding rules on recommendation algorithms, deepfakes and generative AI before most Western governments, published a national AI Safety Governance Framework, and its scientists have signed international statements calling for AI “red lines.”
China's rules do not already prohibit ASI. The premise that China is uninterested in any rules is empirically weak, and a weak premise is a bad reason to race. Filter to China above to see the record.
This is a curated list of the most significant ASI-governance milestones, not an exhaustive legal database. The point of listing them is to show how much activity still leaves the central act unregulated: building superintelligence.
Governance moves fast, if you spot something we've missed or that needs updating, tell us.