BLOOMINGTON, Ind. — A proposed Bloomington ordinance would prohibit distributing or deploying artificial superintelligence in the city.
The Common Council could lift the ban only after making specified safety findings and receiving approval from more than 95% of registered city voters, the proposal says.
Some experts have warned of the risk of catastrophic outcomes from artificial superintelligence, or ASI, which refers to a model that can learn and reason like a human, a step beyond today’s artificial intelligence models.
Bloomington’s proposal would ban autonomous systems that can modify their own core algorithms, run code outside their initial sandbox or perform complex tasks without continuous human oversight if those tasks could affect municipal networks, banking, financial markets or public safety.
A staff memo recommends approving the proposal. The memo says neither Indiana nor the federal government has enacted safety rules for artificial superintelligence. It describes the ordinance as a temporary local measure amid what it calls state and federal inaction. The memo says the measure would set a model for other cities.
U.S Sen. Bernie Sanders, I-Vermont, and Rep. Greg Casar, D-Texas, in September introduced a federal proposal to ban the development and deployment of ASI and to create a Cabinet position to police the industry. In the same month, a British member of Parliament introduced a proposal to prohibit ASI.
Critics — including U.S. President Donald Trump, who has expressed a desire to avoid international regulations — have argued that a domestic ban or slowdown would hand a definitive technological and military advantage to China.
The Bloomington Common Council will consider a second reading of the proposal Wednesday at 6:30 p.m. in Council Chambers at City Hall. People also may attend remotely via Zoom.
