Digital Liberty Fights for the Free Market in Artificial Intelligence
By Rohan Naval
On September 15th, Digital Liberty submitted comments to the Department of Justice in response to an inquiry on state laws that have significant effects on the national economy or on interstate commerce. Digital Liberty used this opportunity to call for federal pre-emption of state laws that regulate artificial intelligence.
The Founding Fathers wrote the Commerce Clause to prevent the chaos of the Articles of Confederation. Under our constitution, states are prohibited from imposing tariffs or taxes on interstate trade for exactly this reason. Artificial intelligence is something that is interstate by its very nature, which gives the federal government the right to pre-empt state regulations.
AI is not a product manufactured and sold within a single state. An AI model may be designed by engineers in California, trained on massive datasets using cloud computing infrastructure located in Virginia and Ohio, and deployed to serve customers in every state from Maine to Hawaii. The data, the algorithms, and the services themselves flow seamlessly across state lines.
This concern is increasingly prescient as states like Colorado pass laws that discourage innovators. The Colorado Artificial Intelligence Act (CAIA), a law that imposes requirements on “high-risk AI” systems to avoid “algorithmic discrimination,” requires companies to abide by the “disparate impact” standard, burdening innovators. Digital Liberty brought attention to the issues that may stem from such a law.
The result is a devastating chilling effect on investment and deployment. Venture capitalists will not invest millions of dollars into a startup knowing that its entire business model could be rendered illegal by the shifting interpretations of a Colorado state regulator. Large, established companies may have the legal teams to attempt compliance, but they will do so by deploying only the most cautious, watered-down versions of their technologies, depriving Coloradans and, ultimately, all Americans of the full benefits of AI.
Furthermore, these laws also put the United States at a comparative disadvantage compared to China in the AI race. America’s current approach with a patchwork of regulations leads to a balkanization of rules, creating regulatory minefields for AI developers who need to operate in 50 different markets with different requirements. When you compare this to the Chinese Communist Party, which puts all its weight behind AI with a unified vision, it becomes apparent that America’s current approach will not beat out China in the long term, unless it changes course.
And while we are engaged in this self-inflicted regulatory war, the Chinese Communist Party will be moving with singular purpose. The CCP is not debating 50 different sets of rules. It is pouring state resources into AI development, integrating it across its military and industrial base, and exporting its technology to the developing world. By allowing our domestic market to become balkanized, we are handing them an unearned and potentially insurmountable advantage. We are unilaterally disarming in the most critical technological race of the 21st century.
Digital Liberty urges the Department of Justice to act on states that wish to punish America’s innovators and uphold the Interstate Commerce clause in the Constitution in doing so. The AI race is a race that America must win, and that can only happen by empowering AI developers in the nation, as opposed to forcing them through regulatory obstacle courses.
The full comments submitted to the DOJ can be read here.