Across federal agencies, artificial intelligence and intelligent automation are playing a crucial role in modernizing data management practices.

With millions of historical records to digitize and standardize, Suman Shukla, the data management section head at the Library of Congress, is leading key efforts to overhaul its approach to handling data.

Shukla has spearheaded efforts to improve data governance at the agency, creating a data governance board, establishing a single source of truth through a centralized data warehouse, and launching a business intelligence working group. She also wants to drive a cultural shift at the agency toward data literacy: “We started a data literacy program to make sure people understand what copyright data means to them.”

Shukla believes data and AI hold the potential to revolutionize how the federal government, and her agency in particular, handles the “gold mine” of data she says she is currently sitting on.

“Data is the new gold today,” Shukla said during a recent ATARC panel moderated by Federal News Network Executive Editor Jason Miller and excerpted for Federal Monthly Insights – Intelligent Automation. “We are talking about AI, which is full on all about data. The Library of Congress is the biggest repository of copyright data, so you can see I’m sitting on a gold mine today.” 

AI and data at the Library of Congress

Artificial intelligence is no longer a thing of the future — it has already had a dynamic and transformative impact on the ability to both improve operations and solve complex data challenges. The same is true for federal entities like the Library of Congress, where leaders continue to seek innovative ways to leverage the power of AI and intelligent automation to make work more efficient and free up brain power for tasks that can’t be automated.

On a recent ATARC panel moderated by Federal News Network’s Jason Miller, Shukla recalled implementing a business intelligence tool at the Library of Congress that generated data into a report that previously required a person to manually compile the same data. That BI tool took the process from a timeframe of two months to just half an hour.

Shukla said the AI isn’t “taking your job,” an idea some resistance to the tool’s implementation centered around. She said it’s making your job more efficient: “You’re free to do more other quality work.”

Government leaders and decision makers want to use the technology to increase efficiency in a way that also upholds transparency and trust. Tools using AI and intelligent automation have become nearly paramount to efficient operations, but their implementation has not been without roadblocks and challenges.

One of the Library’s significant challenges has been managing the massive volume of legacy data, including handwritten records in physical paper format. In digitizing these materials, optical character recognition technology “doesn’t help,” according to Shukla.

“We need some kind of large language processing models or some kind of technology that can extract data, that can look into a human handwriting’s done over a period of time with different policy and procedures,” Shukla said.

With the Library focusing on the need to standardize and digitize data, other agencies are zoning in on different aspects of AI.

Cross-agency AI collaboration at the VA

The focus at the Department of Veterans Affairs is on how AI can facilitate cross-agency collaboration, according to Tony Boese, research programs manager at the VA’s National Artificial Intelligence Institute.

Boese stressed the importance of working together across agencies and with private partners: “We should do it together.… Truly collaborate, not just within and across your agencies, but with interagency collaboration and private-public cooperation.”

Trustworthy AI is a “big deal these days,” Boese said, and requires more than just internal safeguards. It must also be consistent across federal agencies — highlighting the need for AI policy alignment.

“There could be a breakdown and a delay in innovation and in government work or in the general advancement of the country, because we have to go back and recheck and re-ask questions and re-vet something,” Boese said on the ATARC panel.

Boese leads an initiative called Aspire, which aims to use AI to streamline workforce training and cross-agency collaboration. It performs automated assessments to identify knowledge gaps and then provides individualized learning pathways for federal employees.

This ensures employees receive the specific training they need, in a “rigorous, robust way that uses AI but is also secure and interesting in those sorts of ways,” Boese said.

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.