Ryan Steed (on the job market)Jan 31, 2024

Excited to share a new paper with @abebab, Victor Ojewale, Briana Vecchione & Deb Raji

We surveyed 300+ AI audit studies from academia, civil society, govt etc. to understand what work is being done + how it relates to impact & accountability.

https://arxiv.org/abs/2401.14462

AI auditing: The Broken Bus on the Road to AI Accountability

One of the most concrete measures to take towards meaningful AI accountability is to consequentially assess and report the systems' performance and impact. However, the practical nature of the "AI audit" ecosystem is muddled and imprecise, making it difficult to work through various concepts and map out the stakeholders involved in the practice. First, we taxonomize current AI audit practices as completed by regulators, law firms, civil society, journalism, academia, consulting agencies. Next, we assess the impact of audits done by stakeholders within each domain. We find that only a subset of AI audit studies translate to desired accountability outcomes. We thus assess and isolate practices necessary for effective AI audit results, articulating the observed connections between AI audit design, methodology and institutional context on its effectiveness as a meaningful mechanism for accountability.

arXiv.org
AbebaJan 31, 2024
Show threadRyan Steed (on the job market)
I learned a lot from this study — some of the most impactful audit work is done academia (e.g. by journalists or regulators) often using very different methods.
Jan 31, 2024 at 7:02pmWeb
Show threadRyan Steed (on the job market)Jan 31, 2024
We also find that most published audit work focuses products/models/algorithms — so we echo calls from others to expand audit work to include the entire ecosystem of affected stakeholders.