How generative AI can help government address complex problems

Federal agencies face immense pressure to serve the American people by safeguarding health, public safety, national security, and the economy. The faster and more fully that federal leaders understand issues at stake, the more impact these agencies can make.

By identifying the people and groups affected by an issue and considering the factors that might influence behavior, leaders can prepare for various scenarios under which the issue might unfold.

This is what a data analyst calls the problem statement, and getting it right is mission critical. To quote Einstein, “The formulation of a problem is often more essential than its solution, which may be merely a matter of mathematical or experimental skills.”

As an analyst who supports multiple federal agencies that routinely deal with immensely complex issues, I see defining problem statements as an ideal task for generative AI.

Sample problem: The semiconductor supply chain

Let’s walk through an example.

You’ve likely heard about the recent CHIPS and Science Act, which aims to increase domestic manufacturing of semiconductor chips. This might seem straightforward, but once you start digging into the details, it quickly becomes a complex issue with many moving pieces.

Semiconductors require a variety of critical minerals including aluminum, arsenic, cobalt, copper, and rare Earth elements. The lion’s share of these natural resources is mined abroad, raising several challenges and concerns, including:

·       How do we manage our dependence on foreign sources for critical inputs?

·       How do we prepare for and address supply chain disruptions?

·       How do we build domestic production capacity?

·       Who are the stakeholders and how can the U.S. government effectively coordinate their efforts?

If we put a dozen experts in a room together, they could spend weeks expanding this list and identifying stakeholders and variables. But despite their aggregate knowledge and experience, their list might not be exhaustive.

Accelerating federal response to urgent needs

Because generative AI draws from massive databases of human knowledge – or large language models – these tools can identify more variables and possibilities than any cadre of policy analysts or experts. And they can do it in seconds, versus weeks or months, speeding up an agency’s ability to respond to urgent national and global situations.

This provides policymakers with the full context and a more complete understanding of the available – and feasible – options for resolving the issue at hand.

Of course, identifying relevant problem statements is only one step toward addressing major policy concerns. AI-generated problem statements and scenarios demand scrutiny to validate assumptions and to assess and counter the potential for bias. In the field of forecasting, we refer to this as falsification (more on this in a future post.)

As agency missions expand in response to rapidly evolving social, economic, and environmental trends, the challenges they must plan for and respond to will only grow more complex. At Accenture Federal Services, I’m excited to be leading a team that is harnessing the power of AI to better serve our nation’s decisionmakers.

(Check out more about what my colleagues and I are up to at our Generative AI Center of Excellence.)

 By generating complex problem statements and moving them through the falsification process, we’re able to give agencies the information and insights they need – at speed and at scale.