AI Adoption: Factors Agencies Should Consider

The original version of Moore’s law declared that the capacity of an integrated circuit – the backbone of IT – doubles every two years. For decision-makers in the era of artificial intelligence, this could be revised to observe that new versions of AI proliferate at an accelerated rate.

Because AI is so new, yet already has so many offshoots, agencies are wrestling with which type of AI to implement – agentic or generative, for instance – and what the priority challenges are that AI should target. One area where AI offers great potential returns is in identification of waste, fraud and abuse within existing programs. Agencies have a long history of seeking out fraud and waste, which can provide essential information in designing AI systems to accomplish that goal.

Learning Objectives:

• Identify ways to adapt tools that identify fraud in financial fields to identify fraud in your agency’s unique programs, from healthcare to survivor’s benefits to food assistance
• Outline the criteria for evaluating the different forms of AI (machine learning (ML), generative, agentic, etc.) to determine which type best suits your agency’s requirements
• Delineate the role of predictive models to identify data anomalies for further evaluation