Data Governance · 3 min read

NASCIO Survey: 56% of State AI Programs Run Without Dedicated Budgets

NASCIO's survey of 33 states finds 97% have AI policies but only 44% give AI leaders independent standards authority; 56% of programs lack dedicated budgets.

By Priya Raman668 words

Key takeaways

  1. 56% of state AI programs lack dedicated budgets; 75% cite insufficient staffing as a top challenge.
  2. 97% of surveyed states have AI policies, but only 44% of AI leaders hold independent authority to set enterprise standards.
  3. NASCIO published the survey November 17, 2026, covering AI officials in 33 U.S. states and territories.
  4. States with a chief AI officer or equivalent grew from under 20% in 2023 to at least 46 by 2026.
  5. About 75% of respondents use the NIST AI Risk Management Framework.
NASCIO Survey Finds Gaps in AI Governance, Data Readiness Across State Governments - CDO Magazine
NASCIO Survey Finds Gaps in AI Governance, Data Readiness Across State Governments - CDO Magazine — Pexels / pixabay

Fifty-six percent of state AI programs have no dedicated budget, and 75% of the officials running them cite insufficient staffing as a top challenge. Those figures come from a NASCIO survey published November 17, covering AI officials in 33 U.S. states and territories. The National Association of State Chief Information Officers asked how governments are organizing AI leadership, writing governance policy, and preparing employees to use the technology.

The headline tension is structural: states have built oversight scaffolding faster than they have built the authority or funding to enforce it. Nearly all respondents — 97% — said their states have developed AI policies, rules or guidance. Another 83% have processes for agencies to submit proposed AI use cases (specific applications of AI submitted for formal review) before deployment, and roughly 75% use the National Institute of Standards and Technology's AI Risk Management Framework as their reference model.

Who actually holds the authority?

Policy adoption is broad; enforcement power is narrow. Although 72% of respondents have jurisdiction over all executive branch agencies, only 44% hold independent authority to establish enterprise-wide AI policies and standards. Another 16% share that responsibility with another entity. Just over half said they can enforce compliance with AI policies at all.

For data teams, this gap matters more than the headline adoption numbers. An AI governance lead with review processes but no enforcement mandate cannot block a rogue agency procurement or force a data-sharing agreement into place. The survey's own framing acknowledges the constraint: these limits affect the ability of state governments to move AI initiatives beyond policy development and experimentation.

Why the staffing number keeps appearing

Staffing was the most commonly cited obstacle to advancing AI initiatives, ahead of rapidly changing technology, funding, cybersecurity or privacy concerns, and data quality or readiness. The ordering is worth noting for analytics leaders: three of the top five challenges are data-adjacent, and the fifth — data quality — is squarely a CDO problem.

The leadership bench has nonetheless grown fast. In 2023, fewer than 20% of states had a state chief AI officer or someone with a comparable title. By 2026, at least 46 states had identified someone tasked with directing AI, and 25 had an official with AI explicitly in their title. State AI officers expect agentic AI (systems that plan and act autonomously rather than answering single prompts), workforce transformation, scaling AI across agencies, risk management, and AI regulation to consume most of their time over the next three years.

What should data teams take from this?

The report's recommendations read like a governance checklist: give AI leaders clearly defined authority to establish and enforce enterprise standards, invest in dedicated staffing and sustainable funding, strengthen collaboration among technology, privacy and cybersecurity teams, and expand employee training.

For CDOs specifically, the survey underscores that policy progress does not equal production readiness. Moving AI applications into production will require reliable data, effective cross-agency data sharing, and clear accountability for data quality and access — none of which the use-case review processes by themselves deliver.

Three practical implications follow from the numbers:

  • Budgetless programs (56% of the sample) will prioritize policy documents over data infrastructure; expect slow progress on data quality initiatives that lack a named funding line.
  • Where AI authority is shared or absent (60% of states lack independent authority), data governance teams should map their own escalation paths now, before conflicts surface.
  • The NIST AI Risk Management Framework's 75% adoption rate makes it the de facto shared vocabulary; align internal data quality controls to it to stay interoperable across agencies.

The findings also point to the need for closer coordination between AI leadership and data governance teams. As states expand AI use across agencies, data standards, privacy safeguards, risk assessments, and governance processes will determine whether applications can be deployed responsibly and scaled. The survey stops short of prescribing who owns that coordination — which, given that only 44% of AI leaders can set standards independently, may itself be the finding.

Google News: Data Governance (Source)

Price tags

Share this article:

More from Priya Raman

Priya Raman

Show full bio

Correspondent covering regulation, policy and their impact on Analytics Digest is a professional daily at Analytics Digest.

13 articles

Read also

Next article »