admin
Senator Bernie Sanders and Representative Greg Casar just rolled out the “Ban Artificial Superintelligence Act.” It sounds dramatic because it is. The bill would pause “advanced AI,” outlaw anything labeled “superintelligence,” create a new cabinet-level Department of Artificial Intelligence, and threaten prison time and corporate death for people and companies that keep working on these systems. This proposal is the latest shake-up in the national fight over AI — and it raises more questions than answers.
What the Ban Artificial Superintelligence Act actually does
The bill aims to permanently ban systems that it labels “superintelligent” and to pause advanced AI work until a new regulator writes the rules. It includes heavy penalties: up to 20 years behind bars for willful violations and even corporate-dissolution remedies for companies that keep building. The sponsors say the goal is to stop systems humanity can’t control. That sounds noble. The plan reads like someone trying to stop a wildfire by banning matches nationwide — and hoping other countries play along.
Why this plan is risky and unrealistic
There are three big problems. First, the bill depends on a clear, enforceable definition of “superintelligence.” In the real world, intelligence isn’t a single switch you can test in court. Second, enforcement would require global monitoring of chips, compute and labs — the U.S. can’t do that alone. Third, asking China and every other actor to stop at the same time is wishful thinking. If the U.S. ties one hand behind its back while rivals keep training big models, that’s not safety — it’s surrender. Meanwhile, the Ezra Klein interview with Nvidia CEO Jensen Huang showed an alternate view: containment and engineering fixes, not blunt legal bans, are what industry leaders are talking about.
Who’s for it, who’s against it, and the real political math
The bill has picked up strange bedfellows. A faction of the AI-risk crowd, including the Machine Intelligence Research Institute, hailed it as a step toward preventing extinction-level scenarios. Many mainstream researchers called it too blunt and warned it could hand strategic advantage to rivals. Practically speaking, this won’t pass a Republican-run Congress anytime soon. But as a political flag, it signals a wing of the Democratic Party that wants strict limits. That fight will shape hearings, funding and public opinion — even if the law itself never makes it to the statute books.
What conservatives who care about safety and America’s lead should push for
We need smart rules, not panic-driven bans. Focus on export controls, chip-tracking where feasible, targeted safety certifications, and real incentives for safe engineering — plus clear, limited penalties for willful harm. Keep the labs competitive while forcing transparency and audits. If Washington decides to act, make it the kind of legislation that protects Americans and preserves our edge, not a one-size-fits-all ban that reads like political theater. In short: regulate smart, don’t kneecap innovation — and don’t ask China to play referee while we disarm.
