AI Auditing: First Steps Towards the Effective Regulation of Artificial Intelligence Systems
By Edwin A. Farley and Christian R. Lansang
The unprecedented ways in which artificial intelligence (AI) can affect our lives motivate a need for greater regulation of AI. In the face of technological change and legal challenges, auditing of AI has the potential to deliver accountability for the impacts of AI systems. Drawing insights from financial auditing and the current AI auditing landscape, this article identifies reforms necessary to create effective forms of AI auditing. This Note argues that there is a need for government-mandated AI audits conducted by professional auditors following established standards subject to government oversight.
The emergence of an industry of AI auditors and government oversight could deliver accountability for AI without disincentivizing innovation. AI auditing standards should reflect a comprehensive approach targeting three components of AI development: (1) data, (2) model, and (3) deployment. Far from being an increased cost of business, AI auditing and oversight can become a means to accelerate advancement of the technology itself. Over time, as the public gains appreciation of the value of an AI audit alongside expanding audit mandates, a virtuous cycle can emerge, reining in the dangers of AI while advancing the technology in a way that is consistent with its positive potential.
Edwin A. Farley, J.D., Harvard Law School, 2024; Sc.B., Brown University, 2019. Christian R. Lansang, J.D., Harvard Law School, 2022; A.B., Brown University, 2019. Christian is the Head of Growth and Product at Better Life Health.