OpenAI has announced its support for three bipartisan bills related to AI and life sciences currently under discussion in the U.S. Congress, as well as the independent safety assessment provision in the FRONTIER Act. A company spokesperson confirmed this stance to Reuters.
The bills supported by OpenAI are the Web of Biological Data Act, the AI-Ready Bio-Data Standards Act, and the SCALE Biology Act. They also include key provisions from the FRONTIER Act, which requires cutting-edge AI developers to have their safety systems verified by external evaluators.
This is a concrete step forward, going beyond a mere declaration of principles. This is because, following OpenAI's recent call for mandatory AI safety standards applicable across the United States, the company publicly lent its support by specifically naming the bill and its provisions.
| The three bills address the foundation of biological data
The Web of Biological Data Act establishes a single point of access to quality-assured, AI-ready biological data, primarily from national laboratories under the U.S. Department of Energy. The bill outlines a plan to expand this initiative—following an initial pilot program—into a system that meets standards for interoperability, quality, and cybersecurity.
The AI-Ready Bio-Data Standards Act is a bill that mandates the National Institute of Standards and Technology (NIST) to establish definitions, standards, and guidelines for the use of biological data. It focuses on organizing federally funded research data—which varies in format and quality across agencies—so that it can be used for AI training and analysis.
The SCALE Biology Act establishes a biometrology research program at NIST. It can be viewed as a bill aimed at developing measurement methods, calibration standards, procedures, and guidelines to enable comparisons between biotechnology experiments and production results.
| It is both a biosecurity bill and a research infrastructure bill
OpenAI maintains that these bills will help prevent AI models from contributing to the development of biological weapons or synthetic virus threats. The logic is that by determining where high-quality data is stored, who has access to it, and what standards are used to manage it, you can improve traceability and security.
However, the text of the three bills isn’t solely composed of regulations that directly prohibit the use of dangerous AI. They also place a strong emphasis on enhancing the competitiveness of U.S. AI and biotechnology research by ensuring safe access to data and establishing common standards and measurement infrastructure.
This point is crucial. While improved quality and accessibility of biological data can accelerate research in pharmaceuticals, agriculture, and biomanufacturing, data combined with powerful AI requires measures to prevent misuse and control access. Even if the bill is passed, specific security standards and operational guidelines will determine its success or failure.
| The FRONTIER Act aims to institutionalize external verification
The official name of the FRONTIER Act is the Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting Act. It establishes a framework that requires companies developing state-of-the-art models to disclose model information, implement risk management systems, undergo independent audits, and report incidents and conduct ongoing evaluations, based on the company’s size and risk level.
One provision that OpenAI particularly supported is the requirement that leading AI companies use independent evaluators to assess the safety of their models. The goal is not to stop at the developer's internal evaluation, but to have an external organization verify the safety procedures and risk mitigation measures.
For an external evaluation to be effective, the evaluator must have sufficient access to the model and internal records. At the same time, we must carefully address potential conflicts of interest that may arise when development companies bear the evaluation costs, as well as the expertise of the evaluation agencies, the protection of trade secrets, and the scope of result disclosure.
| OpenAI's policy stance has become more specific
In early September 2026, OpenAI argued that nationwide mandatory regulations based on capability levels were needed for state-of-the-art AI. The proposal called for consolidating testing standards, independent evaluations, cybersecurity, and major incident reporting into a single federal framework.
This endorsement is a follow-up measure that links that principle to specific bills. This effectively means they have publicly agreed to move beyond simply having companies voluntarily conduct safety tests and instead establish a common legal obligation.
Nationwide regulations can reduce the costs associated with states operating under different systems. On the other hand, if federal standards are designed too weakly, they could end up preventing stronger state regulations; therefore, the level of the final standards and the enforcement structure are more important than uniformity itself.
| Industry support has a mixed impact on regulatory discussions
Support from OpenAI, one of the leading AI developers, could help the bipartisan bill gain traction in Congress. If companies with technical expertise and evaluation experience participate in drafting the bill, they can also help establish realistic testing procedures.
However, the complex auditing and reporting requirements that large companies can handle may serve as a high barrier to entry for smaller competitors. This is also why the bill seeks to differentiate obligations based on company size.
Support for a policy doesn't necessarily mean the bill will be passed or implemented. During the legislative review process, the criteria for targeted companies, the authority of evaluators, the government's enforcement measures, and the relationship with state laws may change.
| We need to monitor whether the bill passes and whether the evaluation remains truly independent.
The facts that have been confirmed so far are The point is that OpenAI has expressed support for three bills related to biological data and the independent evaluation provision in the FRONTIER Act. It's not as though all of these bills have been definitively passed or that new obligations have been immediately imposed on OpenAI models.
Going forward, we need to verify whether bills pass through standing committees and the plenary session, how independent external evaluation agencies are from developers, and what access controls are in place for sensitive biological data.
This endorsement signals that the discussion on AI safety is shifting from abstract ethical principles to concrete mechanisms such as data management, measurement standards, external audits, and incident reporting. The actual impact will depend not on the endorsement itself, but on the final wording of the bill and its enforceability.
References
Reuters — OpenAI backs bills in US Congress on AI and biological weapon threats
U.S. Representative Lori Trahan — Introduction to the FRONTIER Act
U.S. Senator Todd Young — Introduction of the AI-Ready Bio-Data Standards Act
U.S. Representative April McClain Delaney — Introduction to the SCALE Biology Act
Tags
#OpenAI #AIPolicy #Biosecurity #FRONTIERAct #AI Safety #Independent Evaluation #BiologyData #NIST #U.S. Congress #AI Regulation #Biotechnology #Industrial Policy
