County CIOs on the future of government policy and operations

I was thankful to participate on a panel alongside Cook County, Illinois, CIO Tom Lynch and DeKalb County, Georgia, CIO John Matelski. The panel, titled “Navigating the Future: County Technology and AI Innovation,” was moderated by Mark Lavigne, deputy director at New York State Association of Counties.

While the panel topic looked to technologies of the future, the discussion kept focus on what really matters: The people these technologies are meant to serve.

“How can we do what we do every day, only better, more effectively and more efficiently?” Lavigne said. “How can we serve our employees and our constituents better tomorrow than we did yesterday? AI is a tool that can help us do just that.”

AI is coming, and governments should look at the lessons learned from prior big shifts in tech to define how they leverage AI in new ways.

AI is coming. And changing fast.

Like the cotton mill, electricity, the railroad, or air travel, AI is coming and will be highly disruptive. And those who do not adopt it will be left behind. Yet, at the same time, that doesn’t mean we need to go all in right away, or go in without the proper planning, policies, or investment in people and process.

The rate of technological innovation in AI will far outpace anything we have ever seen. As Granicus Chief Data Officer Bob Ainsbury says, “We shouldn’t try to forecast what will happen,” but rather invest in flexible technologies that benefit from the rapid change of AI innovations.

Your AI strategy should be about more than just tech.

AI will challenge norms and shift the thinking, training, and experiences of government staff. Public servants must leverage AI in various aspects of their job, whether generative, summative, or through query and anomaly detection.

Similarly, processes will need to change to adapt to new governance structures because of AI, requiring tighter connectivity than ever before between policy, operations, and technology.

Yet, we must also acknowledge that change in government is not always easy and is often far slower than we would like.

As John Proffit, Chief Digital Officer for Franklin County, Ohio, wrote in a Substack article, “Why are we so obsessed with AI? We haven’t figured out cloud or human-centered design yet.” This quote got some of the strongest laughs when shared at the NACo Forum, highlighting that driving change in government is hard, and technologies that are long-standing and less volatile than AI are still not mature.

Data — not just for the sake of data — but data specific to government-constituent interaction is key.

Data is key to how LLMs and other machine learning (ML) capabilities will operate. Yet today, those LLMs are informed by open-source data — not data unique to individual counties. And, as Lavigne called out, how will AI work in small counties with just 5,000 residents? Is it just for the large counties that produce millions of interactions?

Technologies offered by Granicus might be a true equalizer in how small and large governments alike benefit from AI. At Granicus, we power some 30 billion (yes, billion) interactions between governments and the public every year. From public meetings to emails and SMS, or digital forms and services that residents interact with like reporting potholes or paying a ticket, this vast data enables us to create highly unique AI capabilities unlike any other in the govtech space.

Different AI constructs have different risks.

We believe there are four common use cases of AI in government that will have the most impact:

1. Generative: Perhaps the most widely discussed AI capability, it’s also the “scariest” to folks given the opportunity for inaccuracies and false or misleading information. It is likely that governments will need to create vastly different policies and processes for how generative AI is used or can be used.
2. Summative: Given governments’ unique requirements around creating and producing public information (e.g. public meeting minutes, etc.) it is likely that summarizing large documents is one of the most significant ways AI can help increase operational efficiency.
3. Anomaly detection and predictive analytics: As governments work to streamline operations, AI is used to identify anomalies or create key efficiencies, time savings, and better manage government resources with predictive analytics.
4. Language translation: AI-assisted translation services will be an important method for ensuring government services are accessible for everyone, including providing multi-lingual broadcasts for public meetings, in-line translation services for public documents, and targeted email and SMS communications translated into the preferred language of the recipient.

AI is not one size fits all. Recognize this and invest in vendors who understand this.

There are thousands of different elected officials, demographics, budgets, and available resources spread across more than 3,000 counties in the U.S. It is unlikely that any two governments will fully adopt the same approach to AI.

As I shared at the forum, customers vary on their approach to AI. Some want to jump in immediately while others are treading more cautiously. No matter how your county chooses to proceed with AI, Granicus can help you find the right solution for your needs.