Federal CIO: Artificial Intelligence Initiatives Will Flop Without Quality Data

ktsdesign/Shutterstock.com

The government’s AI adoption must go hand in hand with improvements to data management, Suzette Kent said.

Artificial intelligence tools are only as good as the data they’re built on, and it’s critical agencies keep that in mind as they ramp up funding for the emerging tech, according to the government’s top technologist.

“If we don’t make the right investments [on] the data side, it’s like we have the fastest plane in the world and no fuel,” said federal Chief Information Officer Suzette Kent on Wednesday.

Artificial intelligence has the potential to improve citizen services, kickstart the economy and create a more efficient government, and AI innovation remains “a priority of this entire administration,” Kent told the crowd at NVIDIA’s GTC DC conference. But AI initiatives will only be successful if agencies first understand the data at their fingertips, she said.

“Getting [data] organized and accessible for AI is no simple task,” she said. “We’re building a foundation for the data in alignment with how we’re enabling [AI] capabilit[ies]. They’re two different journeys, but they have to go together.”

Every day federal agencies collect vast quantities of data on international travel, health care and numerous other areas, but much of it remains unorganized and inaccessible. As agencies struggle to corral that information in a standard, machine-readable format, the Trump administration is looking for ways to help them take advantage of the data at their fingertips.

The White House made improving data management one of the key pillars of the President’s Management Agenda, and its upcoming federal data strategy will include action items for agencies to immediately begin improving their information ecosystem. Officials are also focusing substantial efforts on commercializing that data to spur private sector innovation in artificial intelligence and other emerging technologies.

Beyond better data management, Kent also stressed government’s adoption of AI will hinge on the technological know-how of the federal workforce. Advanced computing tools will “fundamentally change” the nature of work, and it’s important the agencies help their employees adapt to that new environment, Kent said. She added companies and academia must take similar actions to improve data literacy among their employees.

“We’re recognizing that effective and ethical use of AI technologies actually depend on a combination of factors,” Kent said. “It’s very important they’re well synchronized.”

The government is beginning to drastically increase its investment in artificial intelligence as global powers like China try to position themselves as leaders in the technology.

Last month, the Pentagon’s research branch announced it would pour $2 billion into groundbreaking AI research over the next five years, and Kent said her team is putting together a comprehensive picture of AI investment at civilian agencies. She said she will meet with international partners in November to explore different use cases for artificial intelligence in government.