Defense leaders said Tuesday the federal government has entered a new phase of operational deployment with artificial intelligence (AI). Meanwhile, Baltimore officials are deploying targeted pilots to improve public services.
Speaking during a panel discussion at the OpenText Government Summit in Washington, officials agreed that AI is transitioning from a niche capability to a foundational part of government workflows.
“At the department level, we have been shifting from defining some of the strategic guidance on AI to actually implementing that,” said Robert Malpass, deputy chief digital and AI officer for intelligence capabilities at the Pentagon. He added that over the last several months, AI “has gone from something that is an innovation activity … to something that is a component of some of the most important workflows and processes in the department.”
That shift toward operational use is also playing out at the city level, where Baltimore officials are deploying AI to address persistent service challenges.
Rakeim Young, chief of staff for the Baltimore Mayor’s Office of Performance and Innovation, said the city is using AI to move “from reactive service to something that is more proactive and [has] earlier detection,” with initial use cases focused on traffic management and public safety.
One local pilot automates parts of the city’s traffic calming process – previously handled manually – by using AI to assess street conditions and determine appropriate interventions. Another uses camera data to identify illegal dumping by analyzing vehicle movement and object size.
Young said Baltimore’s approach is intentionally incremental. “It’s more important that the AI focus for your organization sits on top of what you already have,” he said, adding that agencies should “start small.”
Despite early successes, officials said core challenges remain consistent across federal and local environments.
Young said the city’s biggest hurdle “wasn’t the technology or the tool,” but gaining buy-in from “city council, the mayor, [and] the general public,” who often associate AI with risk.
Panelists also emphasized the importance of data quality. “The hard work has to be at the beginning whenever it comes to that data governance and that data management,” Caroline Kuharske, chief data and AI officer at August Schell and former chief data officer at the Defense Information Systems Agency, said. Without good data, she warned that AI is “just gonna be garbage in, garbage out.”
At the federal level, those challenges are magnified by scale and complexity. Malpass pointed to the difficulty of integrating AI across agencies and partners, where systems must operate across “a unique tech stack, a different security and authorization process, [and] a different set of legal rules.”
Baltimore faces similar issues on a smaller scale. Young said scaling AI requires coordination across city agencies with “different and disparate data teams,” adding that aligning those efforts “was a challenge” that remains ongoing.
Officials said the path forward hinges on strengthening foundational capabilities, such as data and workforce development.
“People matter,” Malpass said. “No matter how much you’re trying to bring in technology, it still is only possible if you have the right people.”