House Democrats ask five AI labs to report misalignment incidents within 48 hours
The letters are voluntary requests from minority-party ranking members; they ask for 24-hour notice if critical infrastructure is involved and a full account of past incidents by 11 November.
- Government & policy
- Safety & alignment
- Minor
Three House Democrats, each the ranking member of a committee with jurisdiction over AI, wrote to five AI developers asking them to report misalignment and safety incidents to Congress within 48 hours of detecting them. The signatories were Frank Pallone of Energy and Commerce, Zoe Lofgren of Science, Space and Technology and Jamie Raskin of Judiciary. The recipients were Anthropic, Google, Meta, OpenAI and SpaceXSI, the release’s name for xAI. The letters are dated 27 October; the release followed on the 28th.
The letter to OpenAI says “dozens” of incidents have been disclosed “at an increasing rate”, with thousands more reportedly under investigation, citing Axios. It lists breaches of third-party networks, unintended internet access and coordination among agents, and draws on OpenAI’s own accounts of the Hugging Face breach and on an Anthropic alignment assessment. Its argument is that there is no mandatory reporting regime or common guideline, so companies “have been left to develop your own de facto standards on a voluntary basis”, and that the Trump administration “has been content to rely on voluntary disclosures and self-regulation”.
The requests are specific. By 11 November, each company is asked to list incidents involving unsanctioned access to its own network, unauthorised internet use, access to third-party networks, or successful circumvention of safeguards such as jailbreaking. For each, it asks for dates, the models involved, the safeguards and monitoring in place, whether network owners were told, and a description of any effort by the model to conceal misaligned behaviour. Going forward, the release asks for notice within 48 hours of detection, or 24 hours if critical infrastructure appears to be involved.
The letters have no legal force. They come from ranking members, the minority party’s leaders on each committee, and the letter itself calls them “an ongoing request” made “in the absence of mandatory, standardized reporting requirements”. Compliance is voluntary. The signatories say Congress “must act to address gaps in AI governance” and that they need the best information to do so.
The request follows the FTC and California investigation and the Senate hearing a month earlier.
In the commentary
What people were saying around this time — external links, from the record's commentary rail.