The Knowledge Barrier – Research

An honest analysis should identify what it cannot answer. The ICMA 2024 survey establishes where American local government stands on AI adoption. It does not resolve several questions that matter for both future research and future policy.

Does Closing the Knowledge Gap Produce Adoption?

What we know: 77% cite knowledge as the primary barrier. Early evidence from Long Beach and San José suggests literacy programs shift self-reported awareness.

What’s missing: No longitudinal study has demonstrated that communities investing in AI literacy subsequently adopt AI at higher rates, develop more effective governance, or produce better outcomes for residents.

Evidence needed: Longitudinal tracking of communities over time, natural experiments from state-level AI mandates, or randomized controlled trials of literacy interventions with adoption and governance outcome measures.

What Do Workers Actually Experience?

What we know: Roosevelt Institute qualitative research documents work intensification, skill devaluation, and context erosion. MissionSquare surveys show more positive results.

What’s missing: A large-scale study capturing worker experience across all three tiers, in local government specifically, with sufficient sample size to test whether qualitative findings generalize.

Evidence needed: Worker-level surveys across tier spectrum, controlled comparisons of workers with and without pre-deployment training, occupational and organizational context analysis.

Does Sequenced Adoption Reduce Harm?

What we know: The three-tier framework implies that building governance capacity through lower-tier experience before advancing to higher-tier applications is logically sound.

What’s missing: No longitudinal study has compared outcomes across communities that sequenced their adoption against those that did not.

Evidence needed: Tracking communities adopting at different tiers and speeds, with different governance preparations, comparing outcomes over time.

Does Shared Infrastructure Work for AI?

What we know: Evidence from analogous government domains supports shared infrastructure (Feiock’s ICA framework, E911 coordination precedent, GovAI Coalition growth to 700+ members).

What’s missing: Whether demand translates into better governance outcomes. AI may have characteristics making sharing harder: technology velocity, heterogeneous needs, vendor market volatility.

Evidence needed: Prospective evaluation of shared infrastructure initiatives with outcome measures beyond participation—governance quality, service delivery impact, and equity between large and small communities.

What About the Smallest Communities?

What we know: The ICMA survey underrepresents communities below 2,500 population. Open-ended responses suggest at least three categories: those wanting help, those questioning relevance, and those fearing imposition.

What’s missing: Whether the knowledge barrier is the binding constraint, or whether capacity constraints render knowledge investment irrelevant.

Evidence needed: Research specifically designed for small communities, with adapted methods, conducted in partnership with state municipal leagues and regional councils.

How Quickly Will This Analysis Age?

What we know: Different components will age at different rates. Structural findings about knowledge barriers and small-community disparities have historical precedent across multiple technology transitions.

What’s missing: Follow-up data. The ICMA 2024 survey establishes a baseline; its value lies in enabling measurement of future change.

Evidence needed: Repeated survey waves (2026, 2028) to measure whether the knowledge barrier has moved and whether movement correlates with interventions and institutional changes.

Likely Durable vs. Likely to Change

Likely Durable

The knowledge barrier pattern. Knowledge-driven adoption barriers have precedent in e-government, GIS, and body-worn cameras. The specific technology changes; the pattern does not.
The small-community disparity. Resource-driven adoption disparities are structural features of American local government, documented across multiple technology transitions.
The governance-adoption relationship. The correlation between governance infrastructure and AI use (91% with policy vs. 48.5% without) reflects a general pattern.
The knowledge-to-politics ratio. The 8:1 ratio reflects a structural feature of technology adoption documented across domains.

Likely to Require Updating

Specific adoption rates. The 51% non-adoption and 5.4% high-priority figures will change, likely rapidly.
Opportunity gaps. Perceived-potential figures will shift as communities gain experience.
Productivity evidence. Municipal case studies represent early evidence. Independent evaluations will broaden and strengthen the base.
The failure record. New failures will occur. Institutional patterns will likely recur, but specific cases will change.
Tier boundaries. The three-tier framework’s boundaries may shift as AI capabilities evolve.

The ICMA 2024 survey establishes a baseline. Its value lies not in its specific statistics—which will date—but in enabling measurement of future change. The analysis will age. The questions it raises—about knowledge, capacity, equity, governance, and institutional responses to technological change in fragmented democratic systems—will not.