The Massachusetts Institute of Technology (MIT) committee of leaders and scholars have published a set of policy briefs to aid the development of a practical artificial intelligence governance framework for U.S. policymakers. Aiming to promote U.S. leadership in AI, the briefs also seek to limit potential harm from new technology and explore how AI deployment can be beneficial to society. The primary policy paper, “A Framework for U.S. AI Governance: Creating a Safe and Thriving AI Sector” suggests AI tools can likely be regulated by existing US government entities that oversee corresponding domains.
The appointed group of MIT leaders believe it essential to define the purpose of AI tools so that regulations are fit for purpose, therefore pushing a practical approach to AI governance. They propose the use of existing agencies and the extension of legal liability as a means to enforce AI regulations, and emphasize interesting points for consideration like ‘human plus’ legal issues, which take into account the capabilities of AI that surpass human abilities.
The policy brief also touches on the importance of responsiveness and flexibility when it comes to AI regulation. Suggestions such as the advancement in auditing of AI tools, the consideration towards creating a government-approved “self-regulatory organization” (SRO) and a standardized system for auditing AI by an independent organization or federal entity were made.
MIT’s committee also sees value in further research on how AI can be beneficial to society. One paper, “Can We Have a Pro-Worker AI? Choosing a path of machines in service of minds,” by Daron Acemoglu, David Autor, and Simon Johnson, explores AI’s potential to assist workers rather than replace them — a strategy that can promote evenly distributed economic growth.
The policy briefs present various regulatory issues and provide a comprehensive perspective on potential solutions as well as encouraging continued examinations in the field. This commitment to detailed analysis and policy research are crucial steps towards effective governance of AI.
