Public Defender Workload and AI: Empirical Patterns

This article summarizes empirical patterns on public defender workload and AI assistance opportunities.

Background

NCSC data (2024):

• Average public defender caseload: 200+ open cases per attorney
• Recommended maximum: 40-60 cases per attorney
• Some states: 800+ open cases per attorney in misdemeanor dockets

Source: https://www.ncsc.org/

Empirical patterns

NCSC caseload studies (2020-2024):

1. The 50% growth in misdemeanor filings 2019-2023 has not been matched by defender staffing
2. Per-attorney caseloads grew 35% on average since 2019
3. Public defender offices report 30-50% attorney attrition annually
4. Self-reported quality concerns correlate with caseload >100/attorney

AI patterns for caseload management

Triage based on urgency

AI classifies cases by:

• Court date proximity
• Charges and severity
• Available time-to-prepare
• Client vulnerability factors

Task allocation

AI matches tasks to attorneys:

• Continuity (case familiarity)
• Workload balancing
• Specialized training
• Locale preferences

Pre-hearing preparation

AI generates:

• Witness preparation checklists
• Exhibit lists
• Discovery review summaries
• Motion template drafts

Limitations

AI assistance does NOT solve:

• Inadequate staffing
• Funding shortfalls
• Witness cooperation
• Evidentiary complexity

Acknowledgments

This article summarizes public sources including NCSC publications.

Dillon Deutsch has worked with public defender offices on AI caseload management. https://courtgpt.ai