MIT Experts Advocate for Strengthened AI Governance and Regulation

MIT researchers and academics have released an insightful policy paper advocating for the U.S. government to expand its regulatory framework for artificial intelligence (AI). Titled "Creating a Safe and Thriving AI Sector," the 10-page document emphasizes the need to extend existing legal frameworks to encompass AI technologies. This includes proposals for integrating AI into healthcare regulations, particularly regarding AI-assisted diagnoses. Furthermore, the authors advocate for AI to be subject to the same oversight that currently governs governmental operations, such as policing practices, bail determinations, and hiring processes.

The researchers emphasize that any laws regulating human activities should also apply to AI systems. They assert: “If human activity without the use of AI is regulated, then the use of AI should similarly be regulated.” This principle would ensure that the development, deployment, and utilization of AI systems adhere to the same standards as those established for human actions in relevant areas.

The policy paper points to higher-risk applications of AI—which are already being scrutinized under existing laws—as critical points of focus. For example, the regulation of autonomous vehicles, which must meet the same standards as human-operated vehicles, illustrates how existing frameworks can be effectively applied.

The authors also argue that developers of general-purpose AI systems, such as ChatGPT, should be mandated to clearly define the intended purpose of their technologies prior to their release. This accountability would allow regulators to establish specific guidelines for intended uses, ensuring that developers align their systems with these defined parameters.

An important aspect of the paper involves clarifying issues related to intellectual property rights in the context of AI. The authors suggest that mechanisms must be developed to help creators protect against potential infringements, advocating for mandatory labeling of AI-generated content as a potential solution.

The paper highlights the current ambiguities surrounding how existing regulatory and legal frameworks apply when AI is involved. It states: “It is unclear whether and how current regulatory and legal frameworks apply when AI is involved, and whether they are up to the task.” This lack of clarity creates a challenging environment for providers, users, and the general public, who may be uncertain about the risks associated with AI technologies. The authors stress the importance of additional clarity and oversight to encourage the responsible development and use of AI, ultimately maximizing its benefits for all Americans.

According to Dan Huttenlocher, dean of the MIT Schwarzman College of Computing, "As a country, we’re already regulating a lot of relatively high-risk things and providing governance there. We’re not saying that’s sufficient, but let’s start with areas where human activity is already regulated and which society has deemed to be high risk. Approaching AI regulation in this manner is the practical way forward."

This policy paper aligns with several others that explore various facets of AI, including the management of large language models, pro-worker AI strategies, and the labeling of AI-generated content, signaling a broader commitment to establishing a well-regulated AI landscape.