AI Tribune
Wants more rulesUS · StateEUSuperseded

Impose a developer duty of care for catastrophic harms

Make frontier developers legally answerable for catastrophic harms their models enable, reasonable-care standards, attorney-general enforcement, no blanket immunity.

Where it stands

The SB 1047-style liability/duty-of-care model was rejected in California (vetoed Sep 2024) and its successors pivoted to transparency (SB 53; the RAISE Act as chapter-amended). The EU flank was also set back: the AI Liability Directive was withdrawn (announced Feb 2025; formal withdrawal Oct 2025), with BEUC and partners pressing for replacement rules. No jurisdiction currently imposes an SB 1047-style duty of care on frontier developers; the transparency-mandate ask carries the successor vehicles.

first seen 2024-02-07 · last seen 2026-03-01 · locus: developer

Bill-status would miss this

The vehicles it rode

  1. 2024-08-31

    Duty-of-care with AG enforcement passes both chambers

    CA SB 1047
  2. 2024-09-29

    Vetoed

    CA SB 1047
  3. 2025-02-11

    Commission 2025 work programme announces intention to withdraw the AI Liability Directive

    AI Liability Directive
  4. 2025-06-12

    Penalty-bearing version passes NY legislature

    NY RAISE Act
  5. 2025-10-06

    AI Liability Directive formally withdrawn (OJ C/2025/5423); file closed

    AI Liability Directive
  6. 2025-12-19

    Signed with amendments narrowing enforcement

    NY RAISE Act

2 claims · every quote links to its document

The record

For · 0

No on-the-record claims captured on this side. Absence here is a coverage fact, not a finding.

Against · 2

Andreessen HorowitzAgainst2024-06-19
The idea of imposing civil and criminal liability on model developers when downstream users do something bad is so misguided and such a dangerous precedent. First off, the bill requires developers to prove that their models can’t possibly be used for any of the defined hazardous capabilities.
Office of the Governor of CaliforniaAgainst2024-09-29
While well-intentioned, SB 1047 does not take into account whether an AI system is deployed in high-risk environments, involves critical decision-making or the use of sensitive data. Instead, the bill applies stringent standards to even the most basic functions — so long as a large system deploys it.