For the Record: What Do State IT Leaders Think About Emerging Tech?

GT caught up with state technology leaders at the recent National Association of State Chief Information Officers Midyear conference. Here’s what they had to say about artificial intelligence, chatbots, blockchain and other headline-grabbing technologies.

ARTIFICIAL INTELLIGENCE

Amanda Crawford, Texas: We use AI today. We certainly use robotic process automation in a variety of applications and projects across the state. One of the exciting areas for us is using AI for security. That’s certainly a place that we have really seen a lot of gains and tremendous value from having that technology when we’re protecting our most critical systems.

John MacMillan, Pennsylvania: We have AI, we have conversational platforms in many areas, but not every area. And like many other things, it takes people, time and money to create the projects and execute the projects that make those things happen. We’re big supporters of it, but we want to make sure that when a business area is taking advantage of that technology that they understand they have a role in feeding it the intents and feeding it the responses and making sure it’s delivering the business results intended.

Bill Smith, Alaska: We’ve used it on the financial side, we’ve used it on some timecard applications and things like that to see where the value would be. And I anticipate we’re going to be bringing those technologies into our environment in the relatively near future on a small scale, focused on specific problems we’re trying to solve. We really are focused on our foundational environment and getting that architecture put together and so I haven’t seen a huge case for large-scale implementation just yet that makes sense for us.

Bill Vajda, Wyoming: In order to put in place an effective AI you have to have the underlying semantics, the ontologies, the data dictionaries … you have to have all those tuned for whatever smart engagement you want to have. There’s a lot of customization in commercial technology like Google or Microsoft or Amazon that tracks patterns of behavior over time, and they have some internal semantics and ontologies that support their customer sets. But until we get to a point where we have a common understanding, an interpretation of things as simple as fire — if you talk to a fireman, those ontologies and semantics mean one thing; if you talk to an HR specialist, those ontologies and semantics mean something completely different. How fire is associated in Alabama culture might be very different from how fire is associated in Wyoming culture. And there’s not standardization of that at the government level because there can’t be, they’re cultural things. So we have the basic tool sets that we’re playing with now, and we have some really good work that’s been done over the last 20 years to bring us to where we are now. But I think we’re miles away from having an effective AI capability for most of the government applications we’re using.

BLOCKCHAIN

Alan Fuller, Utah: I had a legislator give me a call and say, “I want you to build me some blockchain.” And I was like, I’m not sure what that means, “build me some blockchain.” I think distributed ledger technology has a place. I know it’s heavily used in cryptocurrency, and I look forward to other use cases for it — but I don’t think it really makes sense to say “build me some blockchain.”

Stephanie Dedmon, Tennessee: For us in Tennessee, we haven’t yet seen a use case that makes blockchain the right thing to do. It may exist, and we may a year or two down the road have a different story about how we feel about blockchain.

Steve Nichols, Georgia: For government blockchain use cases, I’m still just not seeing any compelling examples out there. I’ve spent a lot of time looking around and talking to people inside my government who want to roll out blockchain, and that’s always my first question: What unique requirements require this to be blockchain that we can’t solve it with something else? Show me anyone else who’s got something running in production where this was the best solution.

CHATBOTS

Jonathan Askins, Arkansas: We have got a couple chatbots for very specific applications. If we’re going to take chatbots further, we’ve got to make sure we understand the natural language; we’ve got to make sure we understand what our citizens are really asking. I think it can work and I do think it can lift and shift various responsibilities from a help desk and customer service standpoint out to the chatbots, but at the same time, I want to make sure we do that very methodically. If we do it too quickly, we’re going to create a lot of frustration.

Amanda Crawford, Texas: I do think chatbots are getting smarter, but so are we. And I think part of it is a learning process. You have machine learning, but you’ve got human learning too, and we’ve all got to be smarter about that and figure out what’s the best place for chatbots. We’ve had some success with them and as that technology evolves, we’re definitely exploring other ways to implement them throughout government to create a better experience for our constituents.

Alan Fuller, Utah: For me personally, when I use a chatbot, I hate them most of the time. I think chatbots still have a way to go in terms of really meeting the needs of the user or the customer. Most people can tell if they’re speaking to a real person or speaking to a chatbot, and chatbots are often not quite as good, but on the other side of the coin, sometimes chatbots do a really great job at answering specific questions. I’m positive on chatbots and look forward to investing more in them, but I don’t think they’re quite where we want them to be yet.

Steve Nichols, Georgia: We all learned with chatbots what we already knew about websites, which is content is king. Having a framework or the underlying technology is just a small part of it. You really just have to have compelling content. What we’ve learned with chatbots is that you can’t just load up a bunch of questions and walk away. The questions are constantly evolving and that’s a lot of work to constantly go in and tune that stuff up. It just takes people looking at it, what are citizens asking for, what are we not answering, what’s new? During the pandemic it was literally just changing on a weekly basis because no one’s heard of, say, omicron, and suddenly all the questions are about omicron and you’re not ready for it. I think it’s more of a maturation situation of recognizing that the focus was first on the technology and now it’s a whole operational thing where we have to keep up with the content and care and feed this thing as we go forward.

Bill Vajda, Wyoming: I hate them. I think what you’re seeing is that human interaction is critical. People don’t get the immediate comfort they need from chatbots, that they’ve been understood, that their problem is being addressed with the priority that they expect, and that there’s a human being responsible at the end of the day for making sure that their issues are resolved. You have to navigate the menus, you have to know how to interact with the chatbot, you’re not interacting on the terms that you want. So it’s actually a very one-sided communication and for most people that’s not very satisfying. So I don’t think that chatbots are a particularly good solution for government because I think that constituents deserve more.

OTHER EMERGING TECH

Stephanie Dedmon, Tennessee: We are primarily focused on robotic process automation. We have stood up an enterprise capability and are working across agencies to identify potential automations and then developing those. Granted there’s varying flavors of emerging technology, but for us, starting with robotic process automation and then growing from there is our plan.

I do wish we would balance our focus between fundamental principles and emerging tech. On the cybersecurity side, 95 to 98 percent of cybersecurity threats are defeated with basic cyber hygiene. It’s not a magic bullet, it’s not, “Hey, there’s a brand-new bot out there that could just solve the world’s problems.” It’s making sure people can identify a phishing email and not click on it, and it’s making sure that passwords are protected. So emerging tech has certainly enabled us to address some of those issues, but I really want to make sure we maintain the context of our foundational requirements. And then use emerging tech where it makes sense to address specific problems.

Alan Fuller, Utah: One of the things we want to do to address technical debt is to move as much as we can to application development platforms of commercial off-the-shelf solutions, particularly SaaS solutions. If we can buy something at a reasonable price off the shelf, specifically a SaaS solution, we’re happy to buy instead of build. Where we do need to build, we want to build on an application development platform, a low-code/no-code platform, because we can reduce our total cost of ownership. It’s easier to support, easier to maintain, easier to enhance. And if we have someone leave, we can bring in another person who understands that platform. The continuity of support is much better for us.

Another one for us that we’re really excited about is robotic process automation. We’re really looking to invest in automating and using tools to try to reduce where there’s a lot of repetitive, manual work that people are doing. We think we can save the state a lot of money in doing that.

Angelo “Tony” Riddick, New York: We move very fast in New York, as everyone knows. I am careful with leading-edge concepts and leading-edge ideas. We want to execute our job by providing our customers with leading-edge technologies, but we’re very careful not to get sucked in to the next latest and greatest that doesn’t add value to the state. We want to make sure that technology helps add value to the state, helps us deliver services to our residents.

Josh Spence, West Virginia: I think the issue around emerging tech is the concept of the buzzword. It’s not that the buzzword is the problem. It’s that we confuse the buzzword as the goal. When we’re talking about emerging tech and the buzzword associated with it, it’s describing a tool, whether it’s a framework, a service, an appliance, that’s a tool to accomplish a goal. It’s not so much that you want to achieve the buzzword or achieve that as the goal; it’s making sure you understand the business outcomes that you’re wanting to leverage that tool for.