One AI Policy, Dozens of Professions: New Jersey’s Model Governance Experiment

In this blog entry, authors Jason M. Loring, Michelle Ramsden & Andrew R. Lee (Jones & Walker) explore New Jersey’s proposed Assembly Bill A4731, which would require the Division of Consumer Affairs to develop a model generative AI policy for professions and occupations regulated by state licensing boards. The proposal aims to establish a common framework for AI governance across professions, including medicine, engineering, accounting, pharmacy, and psychology. However, the bill has not yet passed either legislative chamber.

The authors highlight the central challenge of applying a single policy across professions with vastly different responsibilities, expertise, and risk profiles. While a shared framework could establish baseline requirements for AI tool approval, data protection, documentation, disclosure, verification, supervision, and incident reporting, individual licensing boards would still need to determine how those safeguards apply to their specific professional obligations and circumstances.

Drawing comparisons with California, Florida, and existing legal profession guidance in New Jersey and Georgia, the authors examine different approaches to regulating AI-assisted professional work. These examples demonstrate the importance of translating existing professional duties into practical requirements, particularly concerning confidentiality, accuracy, independent verification, human oversight, and accountability. They also highlight potential inconsistencies in enforcement if individual licensing boards develop substantially different standards.

Finally, the authors argue that A4731 could provide a useful model for other states, provided its implementation goes beyond a generic compliance exercise. If enacted, the effectiveness of the proposal will depend on the Division’s initial policy framework and the willingness of individual boards to make meaningful profession-specific adaptations. Ultimately, a common governance structure can establish essential safeguards, but professional judgment and accountability cannot be delegated to AI or replaced by a policy template…

CONTINUE READING…