APPCityLife's AI Chatbot Could Help Computer-Less Immigrants Access City Services

The team is working on the project for its submission in the 2017 NYC BigApps competition.

by / April 19, 2017
A screenshot of what a conversation between a user and AppCityLife's chatbot might look like.AppCityLife

If cities want to serve immigrants, what they need is empathy — so says Norma Padron and the team at APPCityLife.

The company, which already sells a mobile-friendly platform-as-a-service product to cities, is working on an immigration-focused solution for this year’s NYC BigApps competition. The direction they’re headed, with an April 30 deadline approaching, is to build an artificial intelligence (AI)-based chatbot. And the banner they march under is empathy.

An empathetic approach would be one where a city service can speak directly to an immigrant in their own language; one where it could understand what they’re saying and knows what they’re looking for. That’s the goal.

And this year, NYC BigApps just happens to be better than ever for that concept. That’s because this year, the competition’s winners could get access to the hundreds of LinkNYC kiosks scattered throughout the country’s biggest city. That is, in order to access APPCityLife’s chatbot, they wouldn’t need a computer, or a phone, or even the ability to speak the English language. As long as they can get to a kiosk, they could use the service to get information and possibly connect with people who can help them. That’s on top of building versions that would work for computers and mobile devices.

The idea, according to APPCityLife Chief Operating Officer Lawrence Abeyta, is to create something that makes it easy for anybody to access the resources they need. Toward that end, they’re building the solution using existing AI engines and natural language processors with built-in translation capabilities. The data they show to users will come from an expanding catalog of open data APIs; if the city opens up more data sets, or if third parties want to offer their own APIs, they could keep expanding the apps' capabilities.

“You could say, ‘I need a doctor’ or ‘I need a lawyer,’ and it could browse across the Internet or across data sets,” Abeyta said.

Existing translation services with open APIs — of which there are multiple choices — have achieved high enough accuracy rates that they should be good enough to serve the average person, he said.

The team’s approach is built from a foundation that acknowledges the challenges immigrants have when trying to use government services, according to Padron. Padron, the associate director of Main Line Health’s Population Health Research Center and a professor at Thomas Jefferson University (TJU) in Philadelphia, knows a thing or two about that — she was born to Mexican parents in a high-poverty Texas border town and has spent parts of her life in three countries besides the U.S. Amid news stories of immigration agents using misleading tactics to find and deport immigrants, she’s witnessed a rising distrust of government that’s seeped into many corners of life.

“Patients [at TJU’s hospital] were very afraid of seeing a doctor and giving them their real names … even in the face of having a heart attack,” Padron said.

She’s drawn on that experience while working with APPCityLife on the project.

“I’m imagining that I’m looking for legal services and I’m in an emergency,” she said.

Still, immigration is just the use case APPCityLife is focusing on for its NYC BigApps submission. Lawrence Abeyta, as well as CEO Lisa Abeyta, see some very natural paths for expanding the idea after the fact.

“Whether it’s someone who’s visually impaired or has problems with language and sound, it gives them the ability to get services when and where and how they need,” Lisa Abeyta said.

As they build, test and deploy the service, they would want to continue improving it. The chatbot could act as a natural data collection service — APPCityLife’s usual approach is to collect opt-in data on an anonymized basis — that could show what kinds of things people use it for the most. It could also generate geotags so cities could see where requests are most likely to come from.

And of course, there’s no guarantee that they win NYC BigApps. But the idea already has government officials’ attention.

“We’ve already started talking with one city who said, ‘When this is ready, we want to be your first,’” Lisa Abeyta said.

Ben Miller Staff Writer

Ben Miller is the business beat staff writer for Government Technology. His reporting experience includes breaking news, business, community features and technical subjects. He holds a Bachelor’s degree in journalism from the Reynolds School of Journalism at the University of Nevada, Reno, and lives in Sacramento, Calif.