Sanders, Casar Announce Ban Artificial Superintelligence Act

This is a maximalist political signal, not yet an actionable compliance regime: the bill’s full text has not been released and its support remains unclear. Axios reports

• –Definitions based on human-level capability, shutdown resistance, or government overthrow could be extremely difficult to measure and enforce.
• –A new cabinet-level regulator would monitor frontier systems, supervise capability removal, and oversee destruction of prohibited systems.
• –Criminal penalties and a corporate “death penalty” would radically raise personal and organizational liability for frontier AI development.
• –International coordination and export controls may limit jurisdictional arbitrage, but verifying compliance globally would be formidable.
• –Developers should watch the bill’s eventual definitions, research exemptions, evaluation requirements, and enforcement thresholds before treating it as an immediate operational constraint.