Ep. 244 Unlocking Federal Efficiency: Cutting Costs and Boosting Security in LLM Development

Connect to John Gilroy on LinkedIn   https://www.linkedin.com/in/john-gilroy/

Want to listen to other episodes? www.Federaltechpodcast.com

Everyone is trying to figure out how to apply AI to federal problems—essentially, building large language models and trying to wring value from them. Inevitably, many are jumping into creating LLMs from various data stores.

We are right at the point where consideration is given to managing enormous data sets in the federal government, emphasizing the need for operational efficiency and security.

The hard lesson learned is data in transit, which means expense.

Today, we will sit down with Dr. Ellison Anne Willimas to explore the potential of privacy-enhancing technologies to enable secure and efficient data use across classification boundaries and data silos.

Dr. Ellison Anne Williams suggests a solution called Privacy Enhancing Technology (PET). It is applied to data as it sits in a silo, a data lake, or whatever nomenclature is used to describe large data sets these days.

PETs allow the secure and private use of data across boundaries and classifications. She explains how PETs enable AI and machine learning models to be trained and used without compromising sensitive data.

The conversation also touches on the cost savings from avoiding data replication and the potential for significant operational efficiencies.

Explore the potential of privacy-enhancing technologies to enable secure and efficient data use across classification boundaries and data silos.