NASCIO CONTACT
Primary Author: Amy Glasscock, Director, Innovation and Emerging Issues
Contributing Author: Kalea Young-Gibson, Senior Policy Analyst
As artificial intelligence becomes a central priority for state governments, leaders must determine how to effectively govern, resource and scale AI initiatives. NASCIO’s inaugural 2026 State AI Officer Survey offers insights from 33 states and territories on how state AI leaders are approaching governance, staffing, procurement, workforce readiness and implementation. The findings provide state CIOs, chief AI officers and other technology leaders with a clearer picture of how AI programs are evolving across states, the challenges their peers are facing and the considerations that can inform their own AI strategies and decisions.

When NASCIO first started tracking in 2023, less than 20 percent of states had a state chief artificial intelligence officer (CAIO) (or a similar title). At the time, it was unclear if the new role would be a short-lived fad, narrowly focusing on the latest technology, or if this short list was a sign of things to come.
Just three years later, at least 46 states have identified at least one person tasked with directing artificial intelligence (AI), and by our count 25 states have someone with AI in their title. It seems the role is growing as fast as the technology itself. This year NASCIO saw artificial intelligence move to the number one spot on the 2026 State CIO Top Ten Priorities List for the first time. States are also facing a host of both possibilities and safety risks as agentic AI and frontier AI models capture headlines.
Given the role’s rapid growth, NASCIO conducted this inaugural survey of state AI officers and directors to learn not only about the role and those who hold it, but about how state AI programs are structured. Thirty-three (33) states and territories responded to this survey on topics related to AI program maturity, governance, workforce, technology and implementation priorities, and the challenges and opportunities they see in their state. In this report we will refer to respondents as CAIOs even though many have different titles (and often have responsibilities outside of AI).
State AI Officer Role
The titles of survey respondents reflect the variety of approaches states are taking to assign responsibility for artificial intelligence. While some do hold the actual title of state chief AI officer, many do not. Among the respondents, 64 percent have titles that explicitly reference AI, while 36 percent hold positions with broader technology, data or organizational responsibilities. Six respondents have a “chief AI officer” title and seven hold director-level titles explicitly tied to AI.
Reflecting the newness of the role, 76 percent say they are the first person to hold the role as it’s currently defined in their state (though this includes those without AI in their title). Of those whose position is specifically designed to oversee AI initiatives, the largest numbers were hired in 2025 (10) and 2024 (5).
When it comes to creating these positions, state chief information officers (CIOs) are generally not waiting for executive orders or legislation to create the role. The vast majority were created by CIO directive (79 percent) though 13 percent chose “established by executive order” and 13 percent indicated that legislation established the role (respondents were allowed to choose more than one).
This is not surprising considering 94 percent said that they are in the state CIO office / central IT organization. Six percent are in the state data office and another six percent are in the department of administration/management. One respondent also said they are under the governor’s office, and one said they are in the innovation/transformation office. There was some overlap.
AI officers and directors are most likely to report directly to the state CIO (46 percent) though they also report to other roles within the CIO organization such as the chief technology officer (21 percent) or deputy CIO (nine percent). Some other responses included chief data officer (six percent) and chief operating officer (six percent).
Fifteen (15) of the respondents said that their role is considered civil service (classified) and eight said their role is non-civil service (unclassified, exempt or at-will). Ten respondents didn’t answer this question.
What professional experience prepares someone to lead AI initiatives for a state? To better understand the career paths of state chief AI officers, we asked respondents about their professional background and current certifications. The following were the most common answers:
Interestingly, AI-specific certifications are less common among CAIOs than professional certifications among chief privacy officers and chief information security officers, with only 25 percent of respondents holding an AI-related credential. This likely reflects the relative newness of the CAIO role and the limited maturity of AI certification programs compared to those in cybersecurity and privacy, which have been established for decades.
Regardless of titles or reporting structures, we asked respondents if AI is their primary responsibility or one component of a broader role. Forty-two (42) percent said that AI is their primary responsibility, 33 percent said AI is a significant component of a broader role, and 24 percent said AI is just one of several responsibilities. One respondent told us about how the expansion of AI expanded their role as well, “It was (my primary responsibility) when I was hired, but AI is in everything we do, so it has expanded to the deputy CIO role. We are planning on hiring another dedicated CAIO.”
Authority and Enforcement
State CAIOs have varying scopes of authority over AI across the enterprise, with 72 percent having authority over all executive branch agencies. About half of them said they serve in an advisory role across all agencies but do not have formal authority.
We wanted to know if state CAIOs have the authority to establish enterprise AI policies and standards. Forty-four (44) percent of officers have this authority, 16 percent share that authority with another entity and 28 percent said that they only have the ability to advise and make recommendations.
We also asked about the authority to enforce compliance with AI policies and standards. Just over half (53 percent) have enforcement authority. Similar to state accessibility officers and state chief privacy officers, chief AI officers would benefit from having full authority to establish and enforce enterprise AI standards. This would reduce agency and staff burden, increase interagency cooperation and unify agencies in the state.
Budget and Resources
The funding landscape for state technology programs has been highly volatile in recent years, as reflected across technology leadership groups, including AI officers. About 44 percent of respondents said their state has a dedicated AI budget; among them, only four states saw an increase within the last year. Fifty-six (56) percent of respondents do not have a dedicated AI budget.
For the states that reported receiving funding, the most common funding source is a dedicated or special legislative appropriation, followed by a state CIO / central IT budget and innovation or modernization fund. Funding can also be inconsistent: one respondent reported only one-time funding, while another explained, “We had an AI appropriation and use general revenue in the IT budget.”
As state agencies look to AI for efficiency and service improvements, we asked state CAIOs if their state provides centralized funding for agency AI pilots or implementations. A third said they do (31 percent), close to another third (28 percent) said the plan is in development and 41 percent said they do not.
Even though many states don’t provide funding to agencies for AI, they do offer many other types of resources. The top resources respondents reported offering to agencies are the following:
- Enterprise licensing for generative AI tools (97 percent)
- Cloud infrastructure (94 percent)
- Governance/risk assessment support (91 percent)
While these numbers are high, it’s important to remember that the states that answered this survey are the ones who have prioritized hiring AI leads, and therefore these numbers would likely be lower if all states were surveyed.
State AI Program Maturity
Fifty-nine (59) percent of respondents characterize AI governance in their state as hybrid/federated and 31 percent say it’s centralized. Only nine percent chose decentralized. When we asked state CAIOs how they would characterize the maturity of their enterprise AI program, 56 percent said their AI programs are established but still maturing and 28 percent said they are in the process of being developed. Only nine percent said the program is established and operational and six percent are in the early exploratory stage.
Governance
Governance, identified by NASCIO as a key challenge for states adopting AI, has made strides over the last few years. We know from our 2026 State CIO Survey that 78 percent of states reported using an AI governance framework, up from 65 percent in 2025. We asked CAIOs to tell us which frameworks they are using (with the ability to select all that apply). Seventy-five (75) percent report using the NIST AI Risk Management Framework with 31 percent following the companion NIST Generative AI Profile. Many are also following state-developed frameworks (59 percent).
We asked state CAIOs what steps their state had taken to operationalize its AI governance framework and many of the options were popular! The majority (97 percent) have developed policies, rules or guidance, 83 percent have created an AI use-case intake process or checklist, 77 percent are conducting statewide AI training, 73 percent have established an AI governance committee, 70 percent have created an AI inventory of state agency applications, 70 percent have implemented AI risk/impact assessments and 60 percent have developed procurement requirements.
Frontier Models
Frontier AI models (the most advanced AI models, with growing capabilities to identify and analyze cybersecurity vulnerabilities) offer new opportunities for states to proactively assess and strengthen the security of their systems. We asked state AI officers and directors if their state is considering the use of frontier AI models to assess or audit cybersecurity vulnerabilities in state infrastructure, software or services.
Ten percent say they are currently using frontier models while 45 percent say they are actively evaluating or piloting their use and another 23 percent are considering using them in the future.
Specifically, we wanted to know for which purposes states are using or considering frontier AI models. The most popular choice among respondents using or planning to use frontier AI models was software/code vulnerability analysis with 83 percent, followed by vulnerability prioritization with 71 percent. Other common answers were infrastructure or cloud security assessment (58 percent), penetration testing / red teaming (58 percent) and security testing of applications or services (58 percent).
Scaling Challenges
While we have started to see states scaling AI projects in the last year or two, challenges remain. We asked states to choose the single biggest barrier to moving AI projects from pilot to production (a tough choice), and answers were widely distributed, though the most common answer (with 22 percent) was funding, which will be a theme later in this survey. After funding, top answers included data readiness, security/privacy concerns and workforce skills, which have been and continue to be the biggest challenges around AI for states.
When asked what will distinguish states that successfully scale AI from those that struggle, state CAIOs emphasized that success will depend on strong governance, executive leadership and operational readiness. One CAIO told us, “The states that successfully scale AI will be the ones that treat it as an enterprise transformation effort, not a series of technology pilots.”
Respondents also wrote about the importance of developing AI skills across the workforce and fostering an environment where employees are prepared to adopt new technologies responsibly. Cross-functional collaboration and coordination across agencies were similarly viewed as essential to scaling AI beyond individual departments or use cases with one state CAIO saying, “It will take a team of people from across the organization collaborating together, not a single person.”
Collaboration
As AI continues to drive state government technology innovation, CAIOs were asked if they see overlap or ambiguity between their roles and those of other enterprise leaders. Officers described inherent, manageable working relationships with multiple executive technology leaders, including the CIO, chief information security officer (CISO), chief data officer (CDO) and chief privacy officer (CPO), through clear governance and decision rights. Many described overlap but not ambiguity or conflict with one writing, “There is some overlap between the CAIO, the chief privacy officer and the CISO, but we work well together.”
However, overlap and ambiguity across roles and responsibilities is expected as every major technology area in state government uses AI, ranging from individual employee use to AI infrastructure in state agencies. Due to role and agency overlap, some officers described ambiguity over who holds primary responsibility across nine core functions of state government technology.
We asked which organization or official has primary responsibility for each of the following areas:
- Vendor due diligence
- AI risk assessment
- AI procurement review
- AI security
- AI privacy
- Data governance
- AI workforce training
- AI use case approval
- AI incident response
State AI officers operate within a complex network of enterprise leadership responsibilities. Survey results indicate that AI activities are frequently a shared responsibility, with between 28 percent and 50 percent of respondents reporting shared ownership across the nine functions surveyed.
AI officers are most likely to have primary responsibility for AI workforce training, AI use-case approval and AI risk assessment. Meanwhile, other established enterprise leaders continue to play prominent roles in their respective domains. CISOs most frequently lead AI security, CDOs lead data governance and CPOs lead AI privacy.
These findings reinforce the collaborative nature of AI governance. Rather than assuming responsibility for every aspect of AI implementation and oversight, state AI officers often coordinate across existing leadership functions. As AI adoption expands, clearly defined responsibilities and decision-making authority will be essential to ensuring effective governance without duplicating established enterprise functions.
Procurement and Vendors
We asked state CAIOs if they or their office is involved in reviewing procurements involving AI. Responses indicated that it varies in both frequency and the point in the procurement lifecycle. This may be because the procurement office holds primary responsibility for vendor due diligence and AI procurement review for 34 percent of respondents. Half of chief AI officers are always involved in AI procurement reviews, with 31 percent usually involved. The point at which CAIOs are brought into a solicitation often varies (34 percent chose this option), but 22 percent reported they are included before solicitation development, 22 percent during solicitation development and 16 percent during vendor evaluation.
When asked to rank their top five vendor issues from most to least difficult, “pricing and cost predictability” was the top issue, followed by “security” and then “transparency into models.” One respondent explained the complexity of AI procurement stating, “A recurring challenge is managing data egress across integrated AI services. Features such as web search or external APIs can move sensitive state data outside the primary tenant or security boundary, creating security, privacy, retention and contractual concerns.”
Formal AI vendor due-diligence processes are still taking shape: 32 percent have one in place, 26 percent do not, and 42 percent of respondents said their process is in development. One CAIO described a cross-functional review process involving procurement, security, privacy, legal and enterprise architecture staff.
Improving vendor-state relationships is a major topic of conversation heading into the next year across all state government areas as states become increasingly reliant on advanced services that cannot be developed in-house.
Workforce, Training and Culture
Thirty-eight (38) percent of officers said statewide AI training is mandatory, and an equal share provides optional statewide training. Training mandate trends here echo those found among state accessibility officers, where uneven training mandates contribute to lower workforce knowledge and uneven goal attainment. Some states provide more specialized training as needed with one explaining, “We have a basic training that is mandatory, and then additional optional training, and have pre-approved vendor contracts for agency specific training if they would like it.”
Training content leans toward responsible use: appropriate use of generative AI appears in every state that reported training topics, followed by responsible AI principles, data and privacy considerations, and bias and fairness. Intellectual property and procurement considerations are the least common topics. In addition to using national training providers, states also develop in-house programs focused on their enterprise, agency and team functions/goals. “We try to align AI statewide training with our AI policies and standards which provide the needed structure and support,” explained one state CAIO.
Sixty-six (66) percent of officers characterize AI literacy in the state workforce as mixed, while 31 percent describe it as moderately or very low, underscoring national trends of AI developing faster than training and familiarity. One respondent noted that half of their state’s employees are not active AI users despite heavy use among the top fifteen percent of users, and another observed higher adoption in IT teams than in business areas.
Shadow AI is the greatest workforce challenge for just over a third of respondents. With highly available and accessible AI tools in the field, it can be a challenge to understand how and when employees are utilizing AI. As one respondent noted, “Just like anything else, if you don’t define the acceptable use of AI … someone else will. You already have shadow AI, and it will grow until you put boundaries in place.”
Just under one-third of respondents chose “lack of general AI literacy” as a top challenge. One CAIO explained the literacy challenge saying, “Uneven AI literacy contributes to uncertainty, concerns about how work will change and inconsistent confidence in using AI appropriately.”
Demonstrating ROI/Value
We have heard from state technology leaders that it’s often difficult to demonstrate the return on investment (ROI) or value from AI, so we asked state AI leaders to tell us what the greatest challenge is. Respondents frequently pointed to the difficulty of measuring benefits that extend beyond traditional cost savings. While AI may improve employee productivity, service delivery, work quality and operational efficiency, these improvements don’t always translate directly into measurable financial returns. As one respondent put it, “Improved employee morale and enhanced citizen experience are hard to measure.”
Several respondents also said that establishing reliable baselines can be a significant obstacle. With no existing performance metrics or a clear understanding of the time and resources required for a process, states can struggle to demonstrate improvements from AI implementation. “(It’s a challenge) helping agencies with baseline numbers so that after AI is implemented, they are able to see the improvements and efficiencies,” stated one AI leader.
Additional challenges included unpredictable AI pricing models and “tokenomics,” workforce concerns and limited resources to move projects from experimentation into production. However, not all respondents viewed ROI measurement as a major obstacle, with some suggesting that the more pressing challenge is finding the capacity and resources to implement AI projects in the first place. As one respondent put it, “This is an exaggerated challenge. There is huge demand for AI productivity tools, it is largely a question of budget cycles and the changing environment of software-as-a-service.”
What State AI Officers and Directors Need Most
AI is evolving at a dizzying pace, and states are working to keep up. And for state AI officers and directors, the job is growing bigger every day. They are faced with a mixed bag: exciting possibilities, critical safety risks, resistance to change, pressure to change and technology that is advancing more quickly than it can be governed. It’s a lot to handle—and many of our respondents just have AI as part of the job.
So, we asked CAIOs about the top challenges they face in their role. The number one challenge was easily “lack of staff” (75 percent). Directing AI for a state government is a massive job that requires many hands right now. The second most common answer was “rapidly changing technology” with 59 percent followed by “lack of funding” with 53 percent—both of which are likely contributing to the need for, and the lack of adequate staff.
With top challenges in mind, we next asked what these AI officers and directors most need to do their jobs effectively. Again, the top answer was “adequate staffing” with 78 percent, followed by “adequate funding” with 69 percent and “strong cross-functional collaboration” with 50 percent.
Digging into this issue a bit more, we asked respondents to tell us what the single biggest thing is that state leadership could do to make them more effective in their role. The following are the most common themes:
- Funding, resources and staff. As mentioned, AI officers need increased investment in AI programs, ability to hire specialized personnel and dedicated time to focus on AI responsibilities rather than managing them alongside other duties.
“Allow me to do this role full time and not in conjunction with other responsibilities.”
- Governance authority. Respondents emphasized the importance of having clearly defined authority to establish and implement enterprise AI policies and governance frameworks.
“Give me a budget and clear authority.”
- Executive sponsorship. AI officers want greater support from agency and statewide leadership to establish AI as an enterprise priority.
“Provide clear, sustained enterprise-level sponsorship for AI as a strategic priority.”
- Strategic alignment. AI officers highlighted the importance of coordinating priorities across agencies, establishing a common enterprise direction and maintaining regular communication as AI capabilities and challenges evolve.
“Frequent alignment and triage meetings. AI issues come rapidly and from a multitude of sources, typically at least requiring a reassessment of priorities; sometimes requiring strategic directional adjustments.”
Advice from the Trenches
We also asked state AI leaders what advice they would give to a state that is considering establishing a chief AI officer (CAIO) position and where the role should sit organizationally. A handful of respondents said that states should consider whether the CAIO position is necessary or permanent with one saying, “A successful CAIO puts themselves out of a job.” However, most felt the role should be more established. Here are the main themes with some direct quotes.
Give CAIOs clear roles, responsibilities and authority.
“Clearly define the role, especially how it is different from other roles that exist within your organization.”
“A state chief AI officer should have the authority to establish statewide AI governance policies, standards, and guidelines to ensure AI is implemented securely, responsibly, and consistently across state government.”
Ensure collaboration and coordination.
“While technology skills are important for this role, that is not the end of it. This is a role for a collaborator who is able to work with several other leaders in the organization and has the adaptability to navigate the ever-evolving landscape.”
“Make sure they are willing to reach beyond their department/state to partner with an ecosystem of external experts.”
Provide executive support and funding.
“Consider clearly what they would do and make sure this is a position with funding and authority.”
“Make sure they have the seniority in the organization”
Position the role under the state CIO
“Should report directly to the CIO because that ensures tighter integration with IT governance and systems.”
“The role should sit within the central technology organization, reporting to the CIO, with clear enterprise-wide accountability.”
An Evolving Role
Finally, we asked state CAIOs how they expect this role to change over the next three years (and they could only choose one answer). Respondents anticipated a shift toward greater strategic leadership (23 percent), AI implementation and adoption (20 percent) and integration with existing technology and data leadership (20 percent). One respondent said, “It will be more strategic, and risk focused because of agentic AI and other projected evolutions.”
Notably, 13 percent also predicted that the role would eventually become obsolete as AI becomes more integrated into all state technology, suggesting that some see dedicated AI leadership as a transitional function rather than a permanent organizational structure.
When we asked what issues they expect to consume significantly more of their time over the next three years, the top answers were agentic AI (90 percent), workforce transformation (60 percent), scaling AI across agencies (47 percent) and AI risk management (37 percent).
Recommendations
Based on the findings from this inaugural survey of state chief AI officers, NASCIO makes these four recommendations for states to improve their enterprise AI programs:
The Path Ahead
AI is advancing rapidly, and state governments are working toward building the leadership structures, governance practices and workforce capabilities needed to keep pace. In just a few years, the state AI officer has emerged as an important enterprise leadership function, though their responsibilities, authority and organizational placement continue to evolve.
The findings of this inaugural survey reveal both meaningful progress and significant challenges, particularly around staffing, funding and the complexity of coordinating AI across government. As states move from experimentation to broader implementation, success will depend on adopting new technologies, empowering leaders, investing in people and building sustainable enterprise programs. With the right foundation, states can navigate the evolving risks and challenges of AI while harnessing its potential to improve government operations and the services they provide to the public.

