Jan 04 2023

5 Questions State and Local Agencies Can Ask About AI

Artificial intelligence can quickly expand citizen services if governments avoid potential pitfalls.

Artificial intelligence technologies, such as machine learning, speech recognition and computer vision, are becoming common building blocks for applications in the public sector. Here are answers to some common questions.

Can AI improve citizen services?

AI absolutely is a valuable tool. For example, state and local governments have become enthusiastic about putting as much information as possible online, but this is often siloed based on department or government level. AI-based tools such as chatbots using natural language processing can work across individual websites easily. Another example: AI-based computer vision is the perfect complement to “see it, click it, fix it” apps many cities are deploying.

Click the banner below to access customized AI content by becoming an Insider.

Can AI be fair and impartial?

AI should be impartial, but it can also fail that test. AI may only be as good as the data fed into it, so the old adage of “garbage in, garbage out” applies here. IT managers must be careful with data sets used in machine learning and watch for deeper biases that might be built into any off-the-shelf tools. Thorough testing in this area is critical, not only for a properly functioning system but for one that meets the needs of the local community.

EXPLORE: How states augment fraud detection with AI technology.

What are the biggest challenges to incorporating AI?

AI has been around for a while, but it has only been practical in the past few years. This makes it challenging to implement in government settings: A lack of familiarity with AI technologies and legacy culture can stop projects before they get off the ground. IT managers can avoid early blockages with adequate training, and by selecting a starting project with off-the-shelf products that quickly show the benefits of AI.

75%

The percentage of governments planning at least three enterprisewide hyperautomation initiatives to be launched or underway by 2024

Source: Deloitte AI Institute, “The Government & Public Services AI Dossier,” August 2021

What’s a good first AI project for state and local governments?

To get started in AI, look for a project that can make incremental improvements to an existing application supporting a citizen service. Being overly ambitious with early projects drives up costs and increases the risk of failure or other negative consequences. Build on small successes rather than reaching for the stars. By focusing on citizen-facing applications, gain leverage in positive publicity, management interest and potential resource savings.

LEARN MORE: How state and local agencies are automating data analysis.

Should an agency build or buy AI?

Aim to do a little of both. Underlying AI technologies, such as engines and frameworks, are best taken off the shelf to leverage vendor expertise. Wherever possible, however, train AI tools on internal data sets. This creates the best and most customized experience possible for end users and has the benefit of creating real ownership of the application within the IT team.

Illustration by John Lanuza

aaa 1

Register