Senior officials from two Federal agencies said this week that they are increasingly using artificial intelligence (AI) technologies to automate routine tasks, boost efficiency, and improve cybersecurity.
During a GovExec event on Aug. 21, officials from the U.S. Government Publishing Office (GPO) and the Department of State (DOS) said the agencies are using AI for aggregating and synthesizing large amounts of data and to improve threat detection capabilities.
Gharun Lacy, DOS deputy assistant secretary and assistant director of the Diplomatic Security Service for cyber and technology security, said that AI can be used to “aggregate an entire risk posture” that allows the DOS to utilize a “dynamic risk picture” for threat detection and protection.
“We’re trying to step away from just looking at our risk from silos – physical security, technical security, cyber security,” said Lacy. “It’s allowing people to do people’s jobs. I don’t need my analyst writing tickets and pushing information to different bureaus when we can automate that and allow the algorithm to do its work.”
“Only a sliver of that intelligence is actually actionable to us,” he said. “So, from the cyber standpoint, those are primary focuses, making sure that we, one, take what we already do and do it better, and then two, sift down the information that we all are inundated with to make it actionable.”
AI also can be used in combination with human analysis to develop defenses, alerts, and monitoring systems for new types of cyber threats, Lacy explained. He said use of the technology also lowers the technical barrier to allow those with limited technical training to utilize AI in developing and protecting against cyber threats.
Lacy also noted that the DOS Center for Analytics is currently exploring AI tech to streamline the transition process for Foreign Service officers – who change roles every two to three years – by helping newcomers quickly adapt and access aggregate contacts and information from predecessors.
GPO is utilizing computer vision – AI-enabled computers which recognize images and videos – in the everyday manufacturing of passports to inspect security features and quality control, explained GPO Director Hugh Halpern.
GPO is also looking to use AI to handle routine proofreading and formatting tasks for documents like those posted to the Congressional Record and Federal Register, which would allow human proofers to focus on more complex editorial decisions. This, Halpern said, can help mitigate the two largest challenges that GPO is facing – around half of the GPO workforce is eligible to retire within the next two years while the number of bills introduced to Congress has “increased exponentially.”
AI is also better suited to catch errors or security risks posed by personally identifiable information (PII) that humans may not catch in documents, Halpern said.
“A machine is going to be a lot better at trying to pattern match and flag for our folks before something goes out,” he said. “Ultimately, [Congress] can release whatever it is they want to release, but we can at least flag it for them and say are you really sure that you want this information?”
Currently, there are more than 700 disclosed use cases of AI across Federal agencies according to a database on the AI.gov website. The Department of Energy and Department of Health and Human Services have the most, with 178 and 157 respectively reported.
