Advancements in artificial intelligence have blossomed in recent years, leading D.C. businesses to look to capitalize on the push by federal agencies investing in the technology sector, said the chief engineer for intelligence programs at MITRE.

MITRE is a not-for-profit that works to support the federal government and its agencies in research and development. Chartered in the public interest, MITRE acts as a bridge between the government and the private sector.

As the chief engineer for intelligence programs, Chuck Howell has seen first-hand how artificial intelligence is advancing within the government. He shared with What’s Working in Washington some concerns about the methods in which artificial intelligence will be used in industry and governed through regulations.

One of these issues is balancing regulation with autonomy in finding solutions. Medical researchers need to strike a balance between “over-regulation that inhibits advances … or being sort of Wild West and laissez-faire and all kinds of bad stuff can happen,” Howell said.

Another problem is accidentally building bias into AI programs, Howell said.

“No one would expect someone to deliberately bake into a machine-learning algorithm, for example, bias. But there are a lot of examples where, implicitly, accidentally, that’s happened,” Howell said. “How does that get regulated or adjudicated or audited? It’s an open question,” he said.

Despite regulations, AI keeps moving forward with recent examples made evident in the public with machines beating humans in games.

In 1997, when an AI was able to beat a grandmaster in chess, it was assumed that no AI would ever be able to win at the Chinese board game Go. “And of course, Google Deepmind is now the world champion in Go,” Howell said.

These are small examples but the advancements are ramping up quickly and Howell wants to find ways to encourage them inside the government. Many fear that the issue of artificial intelligence could become over politicized and leave innovators walking away from opportunities.

“I’m more concerned about us deciding the juice is not worth the squeeze, and just not pursuing the use of machine learning,” Howell said. “I think that would leave a lot of great social opportunities on the cutting-room floor.”

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