Every year, millions of people who qualify for food assistance never get it. The money exists and they meet the eligibility rules, but they don’t know that the program exists, or if they qualify. For others, the application is too long and complicated.

Public interest technology can fill these gaps between what a policy promises and what a person experiences.

Public interest technology is a combination of technology, data, design, policy, social science, ethics and law brought together to improve people’s lives and serve the common good. Public interest technology has prevented children from being poisoned by lead paint, helped families stay in their homes instead of being evicted, and supported disaster recovery efforts after the Los Angeles fires.

The work is carried out by government agencies, nonprofit organizations, universities, community groups and experts like me who study how to use technology to help solve societal issues.

How a solution arises

Public interest technology starts with identifying the problem and asking two questions: What system is not working for the people it is supposed to serve? And how can it be improved?

The next step involves building a cross-disciplinary team to address the situation and scoping the problem. A caseworker, for example, knows things that no analyst can discover from data. A community member knows the reality on the ground. Data scientists, designers, engineers and policy experts, as well as the people affected, create the technology together.

Once there is an initial design for a solution, the team iteratively develops the system, validates that it functions as intended, evaluates whether its impact is equitable, and field-tests it to reveal errors and measure impact.

Impact is the bottom line. A pilot program that ends in an academic paper or a press release does not help people. Impact means the system works, the community trusts it and uses it, and people’s lives improve.

Who the technology helps

Chicago used to wait until a child’s blood test came back positive for lead before inspecting a home for lead sources. That sequence protects the next child who will live there, but not the current one. A data-driven system that we helped develop tracks housing stock, building permits and health information in order to flag high-risk housing units, allowing the health department to conduct preventive inspections before a child gets poisoned.

Allegheny County, Pennsylvania, has a limited budget to help the 10,000 to 15,000 people who receive eviction notices each year. A machine learning system now identifies the residents most likely to become homeless, so the county’s limited outreach efforts and scarce dollars reach the people who need help the most.

Johnson County, Kansas, linked its jail system, emergency medical service and mental health service so it could identify people cycling between crisis and incarceration. The system helps mental health workers identify and reach out to people at risk before the next arrest.

Some cities are expanding public interest technology in-house. New York just launched the Public Interest Technology Crew, which embeds technology specialists in agencies to solve problems for residents. The first crew, embedded in the Department of Consumer and Worker Protection, is building a portal for New Yorkers to report violations of click-to-cancel subscription laws.

Organizations such as the U.S. Digital Corps, Coding it Forward and Data Science for Social Good (which I co-founded) are creating partnerships between technology specialists and employees in governments and nonprofits. And universities are starting to weave public interest technology into curricula, largely driven by the Public Interest Technology University Network, a consortium of universities, cities and foundations.

What makes a technology a public interest technology is a set of choices: who helps build it, how its success is defined, and who is accountable when it gets things wrong. Made correctly, those choices mean the person who qualifies for food assistance isn’t left waiting for it.

.