Failure Modes
Policy Research & Advocacy

Failure Modes

Documenting systemic gaps in policy implementation and proposing precise, administrative fixes to restore regulatory efficacy.

Browse All Memos Learn More
0
Published Entries
0
Policy Areas
0
Institution Targets
24/7
Editorial Coverage
How It Works

The Methodology

A systematic approach to identifying policy failures and designing targeted interventions.

1

Structural Analysis

Identify "failure modes"—specific points where policy implementation diverges from intent due to misaligned incentives, authority gaps, or technical debt.

2

Procedural Fixes

Design non-partisan, administrative interventions that focus on transparency, accountability loops, and human-in-the-loop oversight mechanisms.

3

Strategic Advocacy

Submit findings to oversight bodies, regulatory agencies, and legislative committees to drive incremental but meaningful reform.

Recent Work

View All
Draft Jan 11, 2026

Silent Exemptions: Accepted Agricultural Practices and Enforcement Gaps

Vague statutory definitions create an unreviewable safe harbor for industrial violations.

  • Vague definitions of accepted agricultural practices prevent effective enforcement
  • Lack of mandatory reporting creates a data vacuum for oversight bodies
Proposed Fix: Establish clear, evidence-based criteria for accepted practices with mandatory biannual reporting.
Latest: Pending submission to Arizona Attorney General
agriculture enforcement transparency
Animal Law Arizona
Submitted Jan 10, 2026

Authority Amplification: Agency Displacement as a Systemic Risk in AI Agent Systems

AI agent systems can displace human agency by executing irreversible actions faster than affected parties can respond or contest.

  • Agency displacement occurs when AI systems act faster than humans can contest
  • Individually compliant agents can produce collectively harmful outcomes
Proposed Fix: Incorporate agency displacement as a risk pattern in the NIST AI RMF, triggering requirements for delay buffers and reversible action design.
Latest: Submitted as Public Comment to NIST
AI Governance NIST
AI Governance Federal