OpenAI backs U.S. bills aimed at AI-driven biological weapons risk

OpenAI is endorsing a package of bipartisan bills before the U.S. Congress that the company says will help stop AI models from accelerating biological weapons and synthetic-virus threats, a company spokesperson told Reuters. The ChatGPT-maker is backing the Web of Biological Data Act, the AI-Ready Bio-Data Standards Act, and the Scale Biology Act, and also supports a FRONTIER Act provision requiring leading AI companies to embed independent evaluators to assess model safety.

POLITICS desk — same-day independent confirmation that a frontier lab is publicly lining up behind Congress’s AIxBio data bills and independent safety evaluation rules.

Named bills, as Reuters has the spokesperson: the Web of Biological Data Act, the AI-Ready Bio-Data Standards Act, and the Scale Biology Act. File those three titles as Reuters / the spokesperson. This desk is not inventing bill numbers, and it is not saying any of them passed.

OpenAI is also backing a key provision of the FRONTIER Act that would require leading AI companies to embed independent evaluators to assess the safety of their models, the spokesperson said, via Reuters. Independent evaluators means outside testers checking whether a lab’s powerful model is safe enough — not the company grading its own homework. File that provision as Reuters / the spokesperson. This desk is not saying the full FRONTIER Act passed, and it is not inventing the rest of that bill.

Politico earlier reported the endorsements; Reuters confirmed them with OpenAI, Reuters said. File that sequence as Reuters’s. This desk did not retrieve a Politico URL as a second primary.

Last week, OpenAI said it was pushing for mandatory national AI safety requirements in the United States, Reuters noted. File that as prior company position via Reuters — not a new Sep 15 product claim, and not a law that landed.

Reuters also notes that concerns about advanced AI have escalated, with heads of top U.S. AI labs including OpenAI calling for a slowdown in the pace of advanced AI development. Those weekend pacing remarks are already filed elsewhere. This filing is Tuesday’s bill-endorsement wire only.

Plain English for the rest of the card: these are proposed laws, not passed law. AIxBio = the overlap of AI and biology. Biological weapons = disease or toxins made to harm people. Synthetic viruses = viruses built or heavily edited in a lab. Independent evaluators = outside testers. FRONTIER Act here = the bill Reuters names for that evaluator rule, not a finished statute.

PRIMARY here: Reuters’ 15 Sep 2026 independent Tier B wire — company spokesperson on the record, not an OpenAI newsroom primary. The endorsement of the Web of Biological Data Act, the AI-Ready Bio-Data Standards Act, and the Scale Biology Act; the FRONTIER Act independent-evaluator provision; Politico-earlier / Reuters-confirmed; and last week’s mandatory national AI safety push are Reuters-attributed. NOT claimed: vote counts, that any bill passed, bill numbers beyond the titles Reuters names, co-sponsors for today’s endorsement, an openai.com newsroom post matching this endorsement, enactment, or that this desk sat in Congress. Distinct from the already-filed mike-johnson-no-ai-moratorium, obama-ai-public-oversight-statement, and china-tc260-ai-safety-framework-3-0.