Liability and Insurance

Using tort, strict liability, and insurance markets to internalise the cost of AI harms.

Governance·Exploring·Last reviewed May 1, 2026

This page is a stub. I’ve marked the territory but haven’t written my views here yet. The headings below are placeholders — the actual beliefs, uncertainties, and evidence are still in my notes. If you want my current take on this topic before it lands here, get in touch.

Where I currently stand

<Headline view: liability is one of the few governance levers that doesn't require a regulator to keep up with the technology; insurance pricing is one of the few signals that can plausibly aggregate diffuse harm. The interesting question is whether these tools can be made to work for low-frequency / high-severity AI risks at all.>

Current beliefs

  • Strict liability for frontier-model harms is more enforceable than negligence-based liability. ~XX%<why>.
  • Insurance markets can price ordinary AI deployment risk; they cannot meaningfully price catastrophic risk. ~XX%<why>.
  • <Claim about joint-and-several liability across the supply chain.> ~XX%<why>.

Uncertainties

  • Does liability move developer behaviour earlier in the pipeline, or does it just price risk after the fact? Why it matters: determines whether liability is a safety lever or only a redress lever.
  • What does proof of causation look like for diffuse algorithmic harms? Why it matters: liability without provable causation is symbolic.

What would update me

  • A successful AI-harm tort case establishing a workable causation standard would change the picture.
  • A clear demonstration that insurance underwriting requirements drove safer engineering practice would strengthen the case for insurance as a lever.

Recent reading

  • <date><title><takeaway>.

Related writing

No essays tagged with this topic yet.

Related regions