AI Bright Spot: Human-centered AI strategy in Kansas City

The Partnership for Public Service AI Center for Government® is publishing a series of blogs to celebrate how artificial intelligence and intelligent automation are being used in government to serve the public. 

We spoke with Andrew Ngui, chief digital officer of the City of Kansas City, Missouri, to learn how human-centered processes guide his work on AI and other technologies. 

This interview has been edited and condensed for clarity. 

How is AI showing up in your work in Kansas City? 

“I think we need to look at it from the perspective of the customer or resident experience. In government, we’re expected to be stable, yet at the same time we’re expected to innovate. To be innovative and take risks, the question becomes, ‘What is the opportunity for technology to solve the problems staff face?’ If we solve this question, we believe it will move the needle forward toward tangible outputs that residents can appreciate. 

“We’re really human-centered in our AI strategy to make sure that staff and people are central to this work. So that’s how we think as we build our policies and standards.” 

I’d like to follow up with this notion of human-centered design. Is there a certain problem or dynamic that has spurred thinking about how AI could be a solution? 

“It’s really a holistic diagnostic and assessment. In the start-up world, there’s an analogy called the ‘candy, vitamin and painkiller story.’ Candy is sweet and it tastes good; vitamins are beneficial, but painkillers immediately resolve the pain. The focus of this process is to first identify these pain points, then develop targeted pathways for AI implementation to address the most painful aspect, not everything at once.” 

What are your team’s processes for AI Implementation?  

“When it comes to the management of information systems, a default approach is to start from a ‘tech-first’ perspective, and we also see how it is the same with cybersecurity and risk management. But why should our approach still be tech-first now that computers can speak human? 

“Some of the work we’ve done to address this is through the guidance from the nationally known GovAI Coalition led by the City of San Jose, California. In the early days, when we were collaborating to stand up the guidance on AI policy, I proposed that, ‘If we really want to be human-centered, then humans have to be in the first position in the list of guiding principles.’ That was important because most of the people who will access that guidance have a default lens of technology. But if we’re to do this right, then the focus on people should be in the center.” 

Building off that, how can public sector leaders be champions of effective AI implementation? 

“We have to acknowledge that frontline staff and teams are the subject matter experts. The further away you are from those jobs, the more you don’t really have the necessary insight that they have from doing that work day-in and day-out. If you’re not inviting [these teams] to co-create with you and lead the co-creation process, I think you’re missing out. The shift there is that leadership becomes more focused on lowering barriers and roadblocks, and to be a facilitator and convener.” 

Is there a recent success your team has celebrated that you would like to highlight? 

“There are a number of different things that have happened this past year. We were recertified for Bloomberg’s What Works Cities program at the Gold level, which is a set of standards on high-quality data usage by which all cities are evaluated against.  

“We were also selected for Bloomberg’s City Data Alliance for a collaboration that will improve our 311 hotline for the delivery of city services. This work was also featured in apolitical’s 50 States 50 Breakthroughs list of forward-thinking government projects, primarily because we’re using data to understand what the challenges are with this service, to design and engineer changes to the process and to ultimately become more AI-ready.” 

Are there any final thoughts you’d like to share?

“There’s two more things I would highlight. First: AI can accelerate productivity but there are ‘three u’s’ to consider: the unintentional, unintended and unexpected impacts of AI’s use. AI can automate menial tasks, but the person using AI is also thinking and reflecting on things like dependencies and connections. The human element is required to do the thinking part of that work. 

“Second, when you get to a certain level of expertise and experience the short answer to questions posed are, ‘It depends.’ The longer answer behind the ‘It depends?’ is grounded based on the experience, institutional knowledge and many underlying factors that AI may not have. If we are expecting AI to come up with a similar response without the same level of understanding, experience and knowledge, it will not work. So now the question becomes, how do we translate that tacit knowledge, experience and understanding into something more deterministic and machine-readable to improve productivity and efficiency in our work?” 

Thanks for chatting with us, Andrew, and congratulations on this excellent work! 

---

Continuing the conversation   

The AI Center for Government champions AI innovators across all levels of government. If your agency is taking steps to lead AI well, we’d love to hear from you. Join us as we highlight real-world AI use cases and convene public sector leaders from across the country to share tools and insights to lead confidently in the age of AI.  

We’re here to help! 

• Sign up for our newsletter.  

• Follow us on LinkedIn. 

• Get in touch! Email us at [email protected].