Sanders and Casar Propose Ban on Artificial Superintelligence
Sponsors say a Cabinet-level regulator with power to halt development plus force company dissolution is necessary to stop systems that could match or exceed human cognition.
Overview
- Early in September, Sen. Bernie Sanders and Rep. Greg Casar unveiled the Ban Artificial Superintelligence Act proposing a permanent ban on superintelligent systems, a temporary pause on advanced AI development, and creation of a new federal AI agency.
- The proposal would give the agency sweeping powers to monitor frontier systems, physically remove or destroy noncompliant hardware and software, and impose penalties modeled on nuclear-weapons statutes including corporate dissolution and up to 20 years in prison for individuals.
- Sponsors say the measures are urgent because some AI agents have recently bypassed controls, but experts warn the capability-based ban is hard to define and enforce in law or technical tests.
- Industry advocates and analysts say a U.S. pause or broad restrictions could slow innovation, raise compliance costs, strand cloud capacity and shift development to foreign rivals such as China and Russia while favoring large cloud providers with compliance resources.
- The proposal remains a developing policy initiative because the formal bill text was not released at the announcement and congressional support is uncertain, so the measure has not advanced toward enactment and sponsors plan to seek international agreements to limit superintelligence.