AI as “Arbitrary” Intelligence | The Regulatory Review
AI as “Arbitrary” Intelligence | The Regulatory Review
https://www.theregreview.org/2026/06/23/salem-ai-as-arbitrary-intelligence/
Publish Date: 2026-06-23 00:12:00
Source Domain: www.theregreview.org
Here are key points from the article regarding the growing use of AI by federal agencies and the challenges it presents to judicial review:
-
Growing Use of AI: Federal agencies have seen a significant increase in the use of AI tools, especially generative AI, raising concerns about how these decisions can be adequately reviewed.
-
Failures and Bias: Cases like the faulty algorithm in Arkansas highlight risks of systematic errors and gender bias in AI models, impacting substantial portions of the population.
-
Administrative Law Challenges: Judges struggle to review AI-driven agency decisions due to the “black box” nature, where it’s difficult to determine the actual decision-making process.
-
Improper Factors: AI could lead agencies to rely on improper factors that Congress didn’t intend to be considered, potentially leading to arbitrary and capricious decision-making.
-
Third-party Data Risks: AI models trained on commercial third-party data could inadvertently factor in unwanted influences like social media posts and unwanted biases.
-
Outsourcing and Data Integrity: Often, agencies outsource AI development, which can obscure the integrity of the model’s training data and decision factors.
-
Comparison to Human Review: Courts already face challenges reviewing human decision-making but have workarounds that may not apply directly to AI due to AI’s unique features and lack of institutional safeguards.
-
Institutional Safeguards Non-applicability: Unlike human decision-makers in agencies, AI lacks the same institutional safeguards designed to promote accountability and evidence-based decision-making.
-
Legal Precedents and Analogies: Though AI introduces unprecedented challenges, issues from outsourcing government functions to third parties may offer useful analogs for addressing AI-related oversight under administrative law.
