Can Superintelligent AI Be Banned? Bloomington Proposes a Local Ban

Summary

Bloomington, Indiana, city council member Dave Rollo is introducing an ordinance on October 14, 2026, that would prohibit anyone from distributing or deploying superintelligent AI within the city. If the Common Council approves it, the law could take effect as early as November 4. The proposal defines superintelligence as an artificial general intelligence with superhuman agency, generality, and intelligence, including systems able to modify their core algorithms, run code beyond their initial sandbox, or operate without continuous human oversight. It also incorporates a definition covering intellect that greatly exceeds human cognitive performance across virtually all relevant domains. The ban would apply even when the system is accessed through a browser while the hardware is located inside city limits. Violations could result in fines of up to $2,500 per day. The article notes that superintelligent AI does not currently exist, and distinguishes the proposal from other suggested safeguards such as kill switches, emergency brakes, democratic safety standards, formal verification, and provably safe design. It also says terminology has become less clear because some officials and technology executives use “super intelligence” or “SI” in ways that may not refer consistently to the same concept.