Last week, TechFreedom filed comments to the National Science Foundation on their request for input on the Development of an Artificial Intelligence Action Plan. Our comments recommend that the plan pursue the following priority policy actions: dispel the myth that AI is unregulated, lay out a light-touch approach to AI regulation grounded in existing laws, promote open-source AI, and uphold free speech under the First Amendment.

“The AI action plan should lay out a light-touch approach to AI regulation, drawing on existing law,” said Andy Jung, TechFreedom’s Associate Counsel. “The AI Action Plan should state up-front that AI is already regulated, dispelling the widespread myth that it isn’t. Rather than creating new rules for AI, the federal government should enforce existing laws, which already address most policy concerns related to AI. The Department of  Justice, Federal Trade Commission, and other federal agencies should use existing authority to police AI practices that harm consumers.”

“This plan should promote open-source AI,” Jung continued. “Open-source AI boasts many benefits. The AI Action Plan should highlight research and strive to educate the public on the benefits of open-source AI—while dispelling myths about open technologies. To that end, the plan should implement across the federal government a marginal risk analysis framework for evaluating open-source AI systems: when assessing the overall risk of open-source models, agencies should use a ‘marginal risk and benefit analysis framework’ to weigh ‘the additional risks and benefits’ of open models ‘compared to those that come from’ closed models or ‘other technologies more generally.’”

“The plan must uphold the First Amendment,” Jung concluded. “The AI Action Plan should protect the free speech rights of Americans and AI-focused firms. Given the vast potential of AI, people are reasonably concerned about the accuracy of AI outputs. The solution is robust competition and allowing firms to adopt clear standards—not infringing on firms’ First Amendment rights. The plan should not attempt to define what is or is not “ideological” or biased. Such definitions are inherently political, content-based, and would only serve to prescribe government-favored outcomes.”

###

Find these comments on our website, and share them on X (formerly Twitter) and Bluesky. We can be reached for comment at media@techfreedom.org. Read our related work, including:

About TechFreedom:

TechFreedom is a nonprofit, nonpartisan technology policy think tank. We work to chart a path forward for policymakers towards a bright future where technology enhances freedom, and freedom enhances technology.

</>