AI for the Public Sector is Incredibly Exciting — and a Little Misunderstood

We’re hearing and reading a lot about artificial intelligence (AI) these days. For that reason, we decided to explore how the technology might be used by public-safety and justice organizations during our fourth annual Conference for Advancing the Public Sector (CAPS). The virtual conference was held the last week of September. (Click here to view all sessions on demand.)

The panel discussion featured the insights of Jake Saur, emergency communications center (ECC) administrator, Arlington County (Virginia) Public Safety Communications and Emergency Management; Bob Cozzie, director, Portland (Oregon) Bureau of Emergency Communications, MCP’s Jim Pingel, vice president of digital justice transformation, and me.

The use cases that already are developing are eye-opening and potentially game changers — and hundreds more will emerge as the technology matures and public-sector organizations learn how best to harness the power of AI. But before I dive into that, it’s important to state a critical truism, which is that AI exists to augment human intelligence, not replace it. More on that in a bit.

Just about every public-sector organization right now is dealing with a staffing shortage on some level and AI can help with that. Consider that about 240 million emergency calls are made to the 911 system annually, according to the Federal Communications Commission (FCC). According to Saur, telecommunicators take three times as many nonemergency calls as emergency calls. People often dial 911 when they don’t know who else to call, to report cats stuck in trees, trees felled by a major storm, potholes, traffic lights that aren’t working, or to inquire about what time the parade starts. All such calls eat up precious time and prevent telecommunicators from doing what they have been trained to do.

Now consider how AI solutions are being used in some places to triage 911 calls and redirect nonemergency calls to other resources, e.g., a 311 line. In some cases, such solutions send the caller a text message that contains a link that will get them to the help they need. The AI solutions perform this triage by picking up on key words that they have been taught to recognize. That’s one of the most interesting aspects of AI — such solutions get smarter as time passes through machine-learning technology, which identifies patterns and anomalies in very large datasets.

By moving nonemergency calls to resources that are better suited to handling them, two important things occur: the public is better served and telecommunicators can focus on the calls that are in their wheelhouse. A corollary benefit is that telecommunicators will be less stressed, which will improve their mental health — a very important consideration given the rigors of their jobs, the environment in which they work, and the life-and-death nature of the emergency calls that they field. Further, AI solutions can analyze enormous amounts of data to determine why a telecommunicator’s call-handling times have been increasing, for instance, which would enable ECC officials to address the cause(s) — this too is a vitally important outcome because in every emergency, seconds matter.

On that note, one of the coolest aspects of some AI solutions that are emerging is the ability to flag when a telecommunicator may need supervisor intervention — the solutions do this by transcribing 911 calls and then recognizing certain “trigger” words that indicate a telecommunicator might be on the verge of a breakdown, enabling a supervisor to step in before it occurs — they even can measure tone of voice. Anything that contributes to better telecommunicator mental health is a good thing — for them, their ECC, and for the public they serve.

Interesting use cases are emerging on the justice side of the ecosystem as well. Chatbots and avatars — some of which replicate the human experience — are serving as virtual counter clerks for less-complex cases involving self-litigants. This frees actual clerks to focus on more complex tasks that require their specific expertise.

Meanwhile, robotic process automation (RPA) solutions are relieving personnel from mundane, repetitive tasks so that they can focus on those that have greater value. For example, such solutions can “read” a document, determine its type, capture the case number, identify the litigants and their attorneys, and submit the document into the court’s case-management system. All of this is done far faster and with much less chance of error than if the document was handled by a human.

As in the ECC environment, these capabilities free court personnel to focus on more complex tasks, which is a better use of their expertise and experience.

However, there is a flip side to AI, as there is with most things. A big challenge is that AI is not well understood right now — and that can be disconcerting. The public wonders whether AI solutions eventually will “take over” government like a science-fiction story that has gone horribly wrong. Public-sector personnel wonder whether AI eventually will cost them their jobs.

The answer to both questions is a resounding “no.” As I wrote above, AI exists to make humans better at what they do, not replace them. However, while AI-driven solutions eventually will be part of every aspect of public-sector operations — they’re no panacea. For example, the reader might be aware of a recent, well-publicized incident where a lawyer used a Generative AI solution to draft a court document — one that cited two court cases that didn’t exist. This is a cautionary tale — humans always should review AI-generated outputs. Further, organizations need to develop governance regarding how AI should be used and under what circumstances — we’re doing that here at MCP right now.

We would love the opportunity to help your organization craft an AI strategy that aligns with your environment and needs, so please reach out. In the meantime, I encourage you to view the AI session we presented during CAPS 2023 (click here to be transported to the on-demand landing page) — I am confident that you will find it time well spent. Finally, a future blog will examine an AI solution that Amazon Web Services has developed for the public sector that shows great promise — so, stay tuned.

Chris Kelly is MCP’s vice president of facilities and operations services. Email him at ChrisKelly@MissionCriticalPartners.com.