US Commerce Department report supports ‘open’ AI models

The U.S. Department of Commerce today released a report endorsing “open” generative AI models like Meta’s Llama 3.1, but recommended the government develop “new capabilities” to monitor those models for potential threats.

The report, authored by the Commerce Department’s National Telecommunications and Information Administration (NTIA), finds that open-source models are making generative AI more accessible to small businesses, researchers, nonprofits, and individual developers. For these reasons, the government should not impose restrictions on access to open models, the report suggests—at least not before examining whether the restrictions could harm the market.

This view echoes recent comments by FTC Chairwoman Lina Khan, who believes that open models can allow more, smaller players to bring their ideas to market and thus promote healthy competition.

“The openness of the largest and most powerful AI systems will impact competition, innovation, and risk in these revolutionary tools,” Alan Davidson, assistant secretary of commerce for communications and information and NTIA administrator, added in a statement. “The NTIA report recognizes the importance of open AI systems and calls for more active monitoring of the risks resulting from the wide availability of model weights for the largest AI models. Government plays a critical role in supporting AI development while building capacity to understand and address new risks.”

The report comes as domestic and foreign regulators consider rules that could restrict or impose new requirements on companies seeking to launch open-architecture models.

California is close to passing SB 1047, a bill that would penalize any company that trains a model using more than 10²⁶ FLOP of computing power is beefing up its cybersecurity and developing a way to “switch off” copies of the model it controls. Abroad, the EU recently finalized compliance deadlines for companies under the AI ​​Act, which imposes new rules on copyright, transparency and AI applications.

Meta has said the EU’s AI policies will prevent it from releasing some open models in the future. And a number of startups and big tech companies have come out against the California law, which they say is too burdensome.

NTIA’s management philosophy is not entirely laissez-faire.

In its report, the NTIA calls on the government to develop a continuing program to gather evidence on the risks and benefits of open models, evaluate that evidence, and act on that assessment, including imposing some restrictions on the availability of models where warranted. Specifically, the report proposes that the government examine the safety of different AI models, support research into risk mitigation, and develop thresholds for “risk-specific” metrics to signal when a policy change may be necessary.

As noted by U.S. Secretary of Commerce Gina Raimondo, these and other steps will be consistent with President Joe Biden’s executive order on artificial intelligence (AI), which calls on government agencies and companies to establish new standards for the creation, deployment, and use of artificial intelligence (AI).

“The Biden-Harris Administration is pulling every lever to maximize the promise of AI while minimizing its risks,” Raimondo said in a press release. “Today’s report provides a roadmap for responsible AI innovation and American leadership by embracing openness and recommending how the U.S. government can prepare and adapt to potential challenges.”