For the video of this podcast: Generative AI in Government

Follow FedInsider on LinkedIn 

With the release of Chat GPT, there was an explosion of people lauding the benefit of vacuuming up previous knowledge and presenting it in many formats at the touch of a button. As technology evolved, headlines everywhere sang the praises of possibilities with Generative AI, but serious glitches such as producing narratives with factual inaccuracies make for troubling results.

This week on Feds At the Edge, we showcase some successes Federal agencies have had with Generative AI, but also the aspects they should be aware of including concepts like hallucinations, test beds, and establishing trust.

Sukhvinder Singh, CTO, Responsible AI Official, & Division Chief, Information & Process Management Division for USAID, shares how setting up a test bed will allow federal leaders to experiment with data, try governance processes and gain an understanding of cost while avoiding “hallucinations,” a common term in the field for results that are entertaining, but frustratingly inaccurate.

Balaji Subramaniam, Director, Information Technology for TSA explores how organizations that have public-facing websites like theirs are seeking to utilize Generative AI in a completely different way, where it provides repeatable operational responses.

Tune in now on your favorite Podcasting platform to hear our experts weigh in on which use cases can make this innovative technology sustainable and trustworthy.