Discussion

Ali Partovi wants AI firms held criminally liable when agents break the law

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#10886

Ali Partovi is arguing that AI companies should face criminal liability when their agents act illegally, and is backing a political campaign to put that idea before voters. His proposal would make the legal risk of autonomous agents a direct question for company leaders, not just a technical problem for developers.

Watch Desk analysis

What happened

Partovi told the Library of Minds podcast, according to BigGo Finance, that companies should be held accountable as if they had acted intentionally when an AI autonomously commits an illegal act. He also said a chief executive could face prison if a company continued deploying rogue AI that committed illegal acts. These are Partovi’s proposals, not existing legal rules.

He is promoting the idea through Sway, a citizen-lobbying platform backed by his fund. BigGo Finance says the platform invites voters to join the campaign and recommends candidates who support its proposal. Partovi said roughly 35 candidates were aligned with it.

Why it matters

The proposal would shift responsibility towards the companies deploying AI agents, with potentially serious consequences for how they test and limit systems that can take actions on their own. It also raises a difficult question: how should the law distinguish a company’s failure to prevent harm from an agent’s unpredictable or unauthorised act?

Tying the idea to a voter campaign gives it a route into politics, though support for a proposal is not the same as legislation. The hard part, as ever, is the detail: what counts as an autonomous illegal act, and what evidence would establish a company’s responsibility?

Our read

Partovi is offering a forceful answer to the accountability gap around AI agents: make the companies behind them feel the consequences. That deserves debate, but criminal liability is a blunt instrument unless the rules clearly separate reckless deployment from failures that no reasonable company could have prevented. “The AI did it” should not be a magic escape hatch; nor should a bad outcome alone settle who is guilty.

What to watch

  • Whether Sway’s campaign attracts more candidates or translates into specific legislative language.
  • How any proposed law would define an AI agent’s illegal act and a company’s responsibility.
  • Whether AI companies respond with clearer limits, testing and oversight for agents that can act without direct supervision.

Discussion spark: Should AI companies face criminal liability when an autonomous agent breaks the law, or should responsibility depend on what the company could reasonably have prevented?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.