May 11 2023
Data Analytics

ETIC 2023: Agencies Find That Generative AI Is Here, Ready or Not

The Department of the Interior was taken by surprise when organizations started using ChatGPT to write grant applications.

Some federal agencies are finding themselves dealing with generative artificial intelligence, such as ChatGPT, due to customer demand, whether they’re ready for it or not.

The Department of the Interior is beginning to look at AI because it will be part of the SAP S/4HANA financial management system it’s implementing by 2024. But ChatGPT took the department “by storm” when vendors started using it to score higher reviews and secure grant funding, said Andrea Brandon, deputy assistant secretary of budget, finance, grants and acquisition for the department.

Generative AI describes deep learning models that generate high-quality text, images and other content based on the data used to train them. Grant applicants are forcing the DOI to take the technology more seriously.

“We don’t have any government policy yet that tells them they can’t use it,” Brandon said, speaking Monday at the ACT-IAC Emerging Technology & Innovation Summit. “So, we’re starting to see an uptick in applications from organizations that have historically been trying and trying to get a grant, and voilà — here’s ChatGPT.”

Those same organizations would likely need to hire a grant writer otherwise, she added. The applications do still require refining after ChatGPT generates them.

The U.S. Patent and Trademark Office went so far as to issue guidelines prohibiting inventions created solely by generative AI without human involvement from receiving patent protection.

Click the banner below to receive Insider content after the conference.

State Department Wants Generative AI Outside Its Network, for Now

Generative AI has been met with general enthusiasm at the State Department, though the sentiment is not universal, said Deputy CIO Laura Williams.

“I’m really excited about it from a knowledge management perspective,” she said. “Imagine if you could lay that on top of our diplomatic reporting trove.”

AI could potentially predict the next foreign conflict or economic crisis and provide data that officials could use to respond, Williams added.

Williams said she doesn’t see the State Department starting with generative AI inside its network until an AI policy is firmly in place. But the technology could improve customer experience by using public-facing data to answer questions from people around the world about how to obtain a visa or from citizens trying to report the birth of a child abroad, she said.

DIVE DEEPER: The federal government’s 'blueprint’ for an AI bill of rights.

How Agencies Can Create a Responsible AI Policy

The Organisation for Economic Co-Operation and Development spearheaded discussion about responsible AI, and many of its guiding principles have been adopted by agencies and incorporated into the Biden administration’s Blueprint for an AI Bill of Rights.

Williams spent a year standing up the State Department’s Bureau of Cyberspace and Digital Policy, which launched in April 2022, and said there are lingering concerns around ensuring AI works for citizens, grows economies and isn’t used for “nefarious,” undemocratic ends such as surveillance.

Until the department addresses those policy details, it’s unlikely generative AI will be allowed inside its network.

“Industry and government have to come together to say, ‘What is AI?’” Williams said. “How are we going to apply these principles of fair use, and how are we going to ensure that bias isn’t baked into the methodologies?”

To learn more about the 2023 ACT-IAC event, visit our conference page, and follow us on Twitter at @FedTechMagazine to see behind-the-scenes moments.

XH4D/Getty Images
Close

Learn from Your Peers

What can you glean about security from other IT pros? Check out new CDW research and insight from our experts.