Still early days for state chief AI officers

Oklahoma became the latest state to hire a chief artificial intelligence officer as governments look to get leaders in place to help them use the technology.

The state unveiled Tai Phan, who was previously the chief technology officer for the Office of Management and Enterprise Services, as its new chief AI and technology officer last month. In a statement marking his elevation to the role, Phan said the state has a “remarkable opportunity to rethink how government operates by bringing forward innovation with trust by design to strengthen our mission, improve services and deliver meaningful, measurable impact for the people of Oklahoma.”

Other states have made similar moves. New York appointed Shreya Amin to its CAIO position in March, while several state officials have shifted into the position from other roles. Texas, for example, appointed Tony Sauerhoff to the role after he served as chief information security officer with the state’s Department of Information Resources. Others, like Nikhil Deshpande in Georgia, have taken the role on as part of an expanded portfolio, while Vermont’s Agency of Digital Services combined its AI and data teams and last year appointed Josiah Raiche to lead them.

The somewhat slow moves speak to the cautious approach states are taking to AI and their leadership on the technology, especially as they still try to assess its impact on their governments and who can lead them on it. And it remains a far cry from former President Joe Biden’s now-rescinded executive order on AI, which mandated that every federal agency hire or designate a CAIO. But there is a lot to do for states as they wrestle with AI.

“I'm committed to leveraging AI to enhance services, facilitate operations, and drive data-informed decisions, ensuring ethical guardrails, equity, and transparency,” Amin said in a statement when she was hired. “Collaborating across agencies and with our communities, we will create meaningful improvements in the lives of our residents while positioning New York as a leader in government AI."

A recent survey from the Boston Consulting Group gives an illustration of how much work lies ahead for states as they decide how to structure their AI leadership. BCG found that while 55% of states have a task force or official responsible for addressing AI’s economic impact and 36% have the same for the workforce impacts, only a small number have reported coordinated plans to act on those impacts.

Meanwhile, 39% said the lack of a cohesive statewide AI strategy is the top internal barrier to adopting the technology.

Steven Mills, a BCG managing director and partner who is also its chief AI ethics officer and global leader of the BCG Center for Digital Government, said an “interesting paradox” now exists in state governments, where the vast majority of leaders say AI is crucial for competitiveness and economic development, but very few believe their states have a plan to address it. That then creates uncertainty and a leadership vacuum, he said.

“It sounds funny, but the fact that there is no governmentwide strategy is the No. 1 thing people cite,” for their uncertainty, Mills said. “The real point behind that is, without that plan, it's hard to coordinate across government resources. This is a topic that you really need to muster all of the government to attack, and without some type of a plan, it's hard to make that happen and to get all the agencies moving in the same direction.”

Having someone as a chief AI officer can help provide that leadership and get agencies rowing in the same direction, Mills said. It can be tricky, as the role varies depending on each government and even each agency, but someone must be in charge even if they have many roles and responsibilities.

“There needs to be somebody within agencies, and many times it ends up being a chief AI officer, but part of their remit is upskilling the workforce,” Mills said. “That's both upskilling in the sense of the average worker and getting them using AI technologies and understanding the basics, as well as the technical workforce, who need to understand the much more technical details, building the systems and overseeing others that build the systems. Somebody needs to have that responsibility.”

Chief AI officers will likely need to play a big role in procurement, too, with government spending on AI tools likely to grow massively in the coming years. Kaye Sklar, senior program manager for content and insights at the nonprofit Open Contracting Partnership, said the real test will be how those officials manage their procurement systems to buy, evaluate and manage the tech successfully and responsibly.

“I mean, in my mind, what makes a good chief AI officer is someone who's really collaborative and who facilitates communication across the government,” Sklar said. “Because one thing that we know from procurement, one challenge of procurement is that it can be treated in a very siloed manner, and a chief AI officer who knows technology and who knows procurement can help bridge that gap and enable better collaboration across the government.”

Mills said while progress has been slow, AI’s evolution should push states into working harder to figure out its impacts.

“States haven't moved as quickly as I would have liked to have seen, but now that they're taking action, I think they'll really try to accelerate that, because there's a realization they need to act,” he said.