Dual-Use Foundation Models with Widely Available Model Weights

Policy recommendations embracing openness in artificial intelligence (AI) while calling for active monitoring of risks in powerful AI models. The National Telecommunications and Information Administration recommends the U.S. government develop new capabilities to monitor for potential risks, but refrain from immediately restricting the wide availability of open model weights in the largest AI systems. “Open-weight” models allow developers to build upon and adapt previous work, broadening AI tools’ availability to small companies, researchers, nonprofits, and individuals. NTIA recommends:

Collect evidence through:

a. Encouraging standards and – if appropriate – compelling audits, disclosures, and transparency for dual-use foundation models (including those without widely available model weights);

b. Supporting and conducting research into the safety, security, and trustworthiness of foundation models and high-risk models, as well as their downstream uses;

c. Supporting external research into the present and future capabilities and limitations of specific dual-use foundation models and risk mitigations; and

d. Developing and maintaining a set of risk portfolios, indicators, and thresholds.

Evaluate evidence through:

a. Assessing the lag time between developers introducing capabilities in leading proprietary models, and those same capabilities being made available in open models;

b. Developing benchmarks and definitions for monitoring and potential action if deemed appropriate; and

c. Maintaining and bolstering federal government expert capabilities in technical, legal, social science, and policy domains to support the evaluation of evidence.

Act on evaluations through actions such as: 

a. Restrictions on access to models; or b. Other risk mitigation measures.

 Keep open the possibility of additional government action.