The Las Vegas health department typically selects at random the restaurants it will inspect. But earlier this year, it tried something new. The agency used a software program to analyze tens of thousands of tweets in order to identify possible food poisonings. The program then connected those tweets to specific restaurants and dispatched inspectors to check for any health violations.
The Las Vegas experiment resulted in citations in 15 percent of inspections compared to just 9 percent when inspections were random. The new approach, which saved the agency time and money, was essentially a form of artificial intelligence, or cognitive computing. Unlike big data and analytics, which work off structured data that takes time to collect and analyze, this program was able to quickly calculate possible health problems by reading unstructured data — words and phrases. That’s an exciting advance in computing capability, and it’s something that experts and state and local CIOs alike see as the next big thing in government technology.
Artificial intelligence has received a ton of publicity already. There’s IBM’s Watson, which famously beat two all-time Jeopardy champions, and then there’s — to a lesser extent — Siri or Cortana on our smartphones. These computers almost seamlessly answer any question posed to them. “A.I. is about making better, more informed decisions, and automating those decisions,” says Daniel Castro, director of the Center of Data Innovation. When it’s used in the right way, he says, A.I. has the potential to make government substantially more efficient and citizen-friendly.
The private sector is already experimenting with a form of A.I. known as “chatbots” to automate and improve customer service. That technology could have plenty of applications in government: Trained call center representatives in social services are in short supply and are costly to train, says Steve Nichols, Georgia’s chief technology officer. He foresees the state Department of Human Services using chatbots to help answer some of the routine questions currently handled by human operators.
Criminal justice and public safety is another area that could benefit from A.I. Following the Boston Marathon bombing in 2013, image recognition software combed through hundreds of thousands of images and videos taken at the detonation site to find clues. The White House Office of Science and Technology Policy has held a series of workshops this year to explore how A.I. could further be used in law enforcement. The possibilities thrown out included having A.I. develop better questions during parole screenings and analyze police body camera video.
Castro thinks A.I. that is embedded in robotics, such as self-driving vehicles, could be a big opportunity for government as well. For example, the city of Columbus, which just won the U.S. Department of Transportation’s $50 million Smart City Challenge, plans to use autonomous vehicles to provide transportation between a neighborhood where unemployment is three times the city average to a nearby jobs center.
Of course, there’s still a long way to go and a lot to work out before A.I. becomes more commonplace in government. For one thing, there’s a lack of skilled analysts who know how to work with A.I. For another, there’s the fear that artificial intelligence will strip away jobs, which has made some hesitant to experiment with it. But as Castro notes, people initially feared bank ATMs too. Eventually, though, they stopped caring “if they didn’t talk to a bank teller.” So why not have a chatbot answer routine customer questions and save the complex queries for humans? “Most people,” Castro says, “will appreciate the better value.”
This article was originally published on Governing.
