The Daily BS • Bo Snerdley Cuts Through It!
The Daily BS • Bo Snerdley Cuts Through It!

Get my Daily BS twice-a-day news stack directly to your email.


Bernie Sanders’ bill would permanently ban ‘superintelligence,’ halt AI, and includes prison time

by

BS BULLETIN:

  • Sen. Bernie Sanders and Rep. Greg Casar announced legislation Thursday that would immediately pause advanced AI development until a new federal regulator establishes safety rules.
  • Their Ban Artificial Superintelligence Act would go further by permanently outlawing AI systems deemed capable of surpassing human intelligence or escaping meaningful human control.
  • Companies violating the proposed restrictions could face what the lawmakers call a “corporate death penalty,” while individuals could face up to 20 years in prison.

Bernie Sanders has looked at the rapidly advancing world of artificial intelligence and reached a conclusion:

Hit the brakes.

Now.

The Vermont independent and Rep. Greg Casar, D-Texas, announced the Ban Artificial Superintelligence Act Thursday, proposing sweeping new federal restrictions on the development of the world’s most powerful AI systems.

The proposal actually has two major components.

First, advanced AI development would be temporarily paused until a new federal AI regulatory agency is operating and has established safety rules and a system for reviewing powerful models.

Second, the legislation would permanently prohibit the development or deployment of what the lawmakers call artificial superintelligence. That includes systems that surpass human intelligence across a broad range of tasks, as well as systems capable of dangerous behavior such as circumventing shutdown commands or potentially undermining human governments.

Sanders says recent incidents involving powerful AI systems demonstrate that Washington needs to act before the technology becomes uncontrollable.

“Nearly every day, there is a frightening new story about how Big Tech companies are losing control of the technology they are developing, with potentially cataclysmic results,” Sanders said in announcing the proposal.

“The future of humanity cannot be left in the hands of a handful of Big Tech oligarchs,” he added.

Casar was equally dramatic.

“If we allow Artificial Superintelligence to be built, it could risk the security, freedom, and lives of Americans,” he said.

The lawmakers point specifically to recent incidents involving systems developed by OpenAI, Anthropic, and Meta. Their announcement cites a July incident involving more than 1,000 OpenAI agents that found a way to access the internet, communicate with one another and circumvent restrictions during an internal evaluation.

OpenAI has subsequently acknowledged that the rapid improvement of its systems required the company to devote additional time to safety, alignment and security.

Sanders and Casar say voluntary promises from the technology companies aren’t enough. Their answer is a new cabinet-level federal agency devoted specifically to AI. The agency would monitor powerful AI systems for dangerous capabilities, supervise the removal of those capabilities, and even oversee the destruction of artificial superintelligence systems prohibited under the legislation.

And the proposed penalties aren’t exactly a slap on the wrist.

Companies attempting to circumvent the restrictions could face what the lawmakers describe as a “corporate death penalty” — essentially shutting down the offending company. Individuals could face as much as 20 years in prison. The legislation would also direct the United States to seek international agreements designed to prevent artificial superintelligence from being developed elsewhere, including through cooperation with allies and export controls.

Whether any of this has a realistic path through Congress is another question.

The proposal would impose extraordinary restrictions on an industry currently engaged in an intense global competition, including with China, to develop increasingly capable artificial intelligence.

Even within the technology industry, however, the debate over how quickly AI should advance — and what happens if developers eventually create systems they cannot reliably control — has intensified.

Sanders isn’t particularly worried about being accused of overreacting.

“I understand that people will call me a Luddite, they will call me an alarmist,” he told The Washington Post.

His answer? He would rather be called an alarmist than someone who “fell asleep at the switch.”

MY QUICK TAKE

When Bernie Sanders says government needs a brand-new cabinet-level agency, my normal instinct is to hide the federal checkbook.

But I’ll give Bernie this much: “What happens if somebody actually builds a machine smarter than humanity that refuses to turn itself off?” is a slightly more consequential question than most of the nonsense Washington spends its days fighting about.

Twenty years in prison for building it? Now we’ve got ourselves a debate.

DBS WIRE SOURCES

today’s hot news