AI continues to excite many — but not all — state IT leaders, report shows

The National Association of State Technology Directors on Tuesday published a report showing that generative artificial intelligence and other types of AI are exciting the imaginations of many state technology officials, though many others remain skeptical.

The report, which collects survey responses from technology officials in 42 states, reveals that most are developing many new AI policies and organizing their information technology departments in preparation of using new AI software in their agencies’ operations. The survey, distributed last May, provides a snapshot of the attitudes and governance strategies found in state governments nearly two years after ChatGPT’s public release.

(The eight states that didn’t participate in the survey are California, Colorado, Hawaii, Idaho, Maryland, New Jersey, New York and Oregon.)

One benchmark for maturity is whether a state has defined the term “AI.” According to the survey, 52% of states have defined AI for the purpose of governance, while 22% have only established a definition and 26% have not defined AI.

Some state IT leaders interviewed by StateScoop have reported excitement with AI’s recent advances, while others have responded with caution or skepticism. One state technology official quoted in the new report said: “In large part, AI has yet to deliver on a lot of what has been promised. We find it important to put a governance framework in place, but a lot of the solutions presented are either not fully functional as expected or utilize things that were already available in large part.”

Another maturity benchmark is a statewide process of cataloging active applications that use AI — 67% reported having conducted such an inventory.

Sixty-two percent of states reported they’re developing preferred contract language for purchases of new AI technologies.

The survey also lends insight into how states are organizing their AI governance. Ninety-five percent of respondents said it was the state chief information officer who is leading AI governance efforts. Thirty-six percent have involved their legislatures, and 29% have included chief information security officers.

Asked about uses of AI they thought were most promising, survey respondents named cybersecurity, citizen portals, data management and analytics, and office worker efficiency. Asked about the most common AI applications already being used, respondents most frequently cited cybersecurity, administrative processes, and transportation and infrastructure.

Top challenges named were the perceived risk of AI, staff skill sets, and budgeting and funding issues.

Forty percent of states reported that they’re teaming up with outside agencies for extra help developing their AI policies and projects. Thirty-three percent reported partnerships with higher education, 19% partnerships with other states, 12% with counties and 12% with cities.

Concerns abound that AI won’t live up to the hype it receives, but many state technology officials report optimism on the technology’s capacity to mitigate tedious work and augment understaffed offices. An emergency communications director using AI to triage 911 calls in New Orleans told StateScoop last July that he expects the software he’s testing will add the equivalent of two full-time personnel to his call center.

In the new report, one official said: “We have more than 40 projects at the idea phase and are awaiting funding, approvals, trained staff and procurement.”