The foundation is set for its AI work; the governor’s office and the legislature have indicated they support AI adoption in accordance with the state’s AI strategy. Heimann’s appointment as South Carolina’s inaugural director of AI is intended to help guide that advancement. However, while Heimann said a lot of agencies in the state are interested in adopting AI tech, many are seeking guidance on beginning to safely harness its potential.
Heimann became director in October, bringing with him a breadth of knowledge in data and AI leadership. His experience includes private-sector roles such as serving as the chief data scientist for Data Tactics and as the chief AI officer for Cybraics Inc. and SilverSky. Last year, he published a book on AI readiness and risk assessment, General Artificial Intelligence Revealed.
Fifteen years working with machine learning and statistical modeling also helped him prepare, the director said. AI adoption failures, he noted, often result from leadership, not technical challenges. South Carolina’s approach, he underlined, aims to balance excitement with pragmatism toward AI adoption, as there are about 60,000 government workers that will be impacted. A key piece of the state’s goal to protect citizens throughout this process is governance, he said.
To achieve its governance goals, the state will launch its COE to assist state agencies in evaluating use cases and providing recommendations to mitigate risk; it will also foster governance communications across agencies. The COE was called for as part of the state’s AI strategy and is slated to launch in January, according to Heimann.
The COE team will be made up of internal experts including Heimann and the state’s chief privacy officer; chief information security officer; and a governance, risk management and compliance expert.
The COE model has become increasingly popular among state governments looking to responsibly advance AI. Utah established such a center in 2018; more recently, Arkansas, Texas and Rhode Island launched similar initiatives.
South Carolina was primarily discussing AI at a conceptual level when Heimann first started in the role, but he said including examples of use cases in the conversations helped stimulate thinking around risk mitigation for those who could be impacted.
“My background is a practitioner’s background,” Heimann said. “So, I think those conversations — like the practical, real-world implications of a specific use case — are much more useful to talk about risk because it’s tangible.”
Heimann said an AI-powered chatbot is an “obvious starting point” for state government to improve citizen service delivery. He said he foresees the opportunity to scale this idea to different agencies to help constituents access information they need or want.
There are often two conflicting perspectives on AI, Heimann said, with one side overstating its risks and the other overstating its usefulness. South Carolina, he said, is more focused on long-term positioning to be better prepared for the technology’s maturation. He applauded the state’s intentional approach, based on the understanding that adopting AI too quickly could present risks.
And while he declined to give specifics, Heimann hinted at forthcoming AI use cases, noting the state’s use of a risk management approach.
