The Future of AI for Public Good

At the close of 2022, artificial intelligence (AI) didn’t crack the National Association of State Chief Information Officers’ list of the “top ten priorities” for the coming year. However, by December 2023 AI took third place, and could well be number one by the end of this year. 

As advances in AI move at seemingly lightning speed, understanding its benefits, challenges and solutions to those in the public sector has become essential.

With that in mind, in September the IBM Center for the Business of Government gathered a group of 30 leaders and experts who work in and with the public sector to discuss a wide swath of issues pertaining to how AI can be implemented ethically and responsibly. The results of these discussions provide an overview of how government uses AI for good outcomes today, and where the technology might be heading in the future. 

This first of four blog posts drawn from that conversation focuses on the benefits of AI. The next three – which will focus on challenges, solutions, and a round-up – will appear over the next three months.

One of the topics that came up repeatedly was AI’s capacity to help translate complex documents into plain and easy to understand English, offering more access to information that might otherwise be indecipherable. 

For example, AI can help people get the information they need from their governments, both for individual transactions and to learn more about the public sector generally. “Having chatbots and things like that, can connect people to services they need,” said one attendee, “in order to make sure that they’re able to be used by everyone.”

One roundtable participant recalled that Minnesota established an Office for Enterprise Translation (ETO) built to remove barriers to access for non-English speakers. As the Office’s website explains, “Established under the mandate of Minnesota Statute 16B.373, the ETO is a centralized and specialized translation services provider dedicated to breaking down language barriers and ensuring that all Minnesotans have equal access to vital information and services. With a focus on Spanish, Hmong, and Somali languages, the ETO also provides translation support for all other languages.”

At the same time, AI can also help governments themselves fully understand the needs of residents. For example, agency staff can easily leverage AI to keep track of potholes, broken streetlights, or food deserts, rather than relying exclusively on self-reporting by citizens through 311 calls or other means. This can be accomplished in a number of ways, such as cameras attached to waste management trucks that observe pothole locations and broken streetlights.

There was broad consensus in the group that AI will not replace government staff, but rather augment their work to serve the public -- used to help provide residents or governments with the assistance they need.   For examples, for any but the simplest queries (like the hours of operation of a DMV), AI has high utility to provide staff at call centers with easy access to accurate answers.

Roundtable participants made it clear that citizens can use AI not just to interact with their governments, but also to ease their way through a variety of services provided by non-government support organizations. As one participant pointed out, “Thirty million people a year go through the system without a lawyer,” but they must deal with complex issues (like) evictions, child custody, divorce and other matters essential to their wellbeing. But natural language processors can help an individual who is going online to provide the guidance they need in lieu of an attorney” -- as noted above, can AI help people make better decisions in providing services to the public.

One of the most powerful uses of AI for government program involves complex grant finding and application processes, by developing systems where “both large and small grantees feel they are being guided through the process,” according to one expert participant. This person went on:

“The prospect of making it easier and more fluid for applying for grants from the federal government for various grantees, and in this case potentially adding in the ability to autofill forms (has been) very attractive to grantees,” said the same attendee. This is particularly helpful for “smaller entities that may not have grant managers, which is certainly true in many, many places.”

The use of AI to help with decision making necessary to help individuals benefit from social programs was also a topic of discussion. One participant cited an example from the University of Albany. According to the University’s website, the school currently works with data sets to “enable early detection of mental health conditions, track patients’ progress, and monitor them after medical intervention, alleviating shortages of in-person professional care where resources are currently scarce.”

A number of other potential benefits from AI also arose during the conversation. The list includes:

• Providing faster call responses for taxpayers.
• Making sure that the findings of administrative representatives, such as health facility inspectors, are supported by state regulation.
• Enabling easier access to data for governments.
• Educating citizens.
• Uncovering discrimination.

All of the above may make it sound like AI can solve any problem that government faces. But if not implemented in an ethical and responsible manner, significant risks may arise. With that in mind, the next installment of this series will deal with challenges that AI presents for government services.