As Federal agencies are increasingly utilizing AI in their processes, it’s important that they approach AI using the mindset of understanding the technology as tools to help workers, rather than technology that will replace workers.
That mindset has been a priority when working AI into operations at the U.S. Patent and Trade Office (USPTO), USPTO CIO Jamie Holcombe said Tuesday at an ACT-IAC virtual event.
“If I had to do a mic drop on this whole discussion, it would be augment not replace,” Holcombe said at the event. “And that’s what we’re saying … for scaling, you have to augment. You’re not replacing. You’re getting rid of the clerical, the administrative, the human being can only go so far.”
Holcombe said tackling that fear of AI-driven replacement, and cultivating that culture change is one of the biggest challenges to implementing AI in the Federal government. The goal of that shift is to get employees to view AI as a tool that allows them to get away from mundane processes and focus on thinking more strategically.
One project USPTO is working on utilizing AI that falls within this “augment, not replace” mindset is underway in an art division at USPTO. Rather than using extremely complicated Boolean searches, the algorithm searches for relevant prior art to simplify assessors’ jobs.
“We have created great machine learning algorithms for different art units which are increasing the speed at which we can build around and create haystacks. The theory of the haystack is finding that needle,” Holcombe said. “Let the artificial intelligence do that – create the haystacks – such that the human can find that needle in that haystack. And that is sort of our conceptual paradigm of the new artificial search algorithms.”
Holcombe said it is important to not only loop human operators in as part of the process but also to be willing to fail. He emphasized failing small, though.
“The other thing that I tried to do is take things in small chunks, and learn in a small way, but you can also fail as one way of learning, failing, and succeeding. Learning includes both,” Holcombe said. “If you don’t fail in your life, you never learn.”
“You have to fail to understand what not to do again, people try to get away from failure so much. Now I’m not saying that you should fail big time,” Holcombe added. “No, you should fail small, so that you learn from the mistakes, and then you can see in the future. And if you keep on … you feel a little here, you grow big there, and you get better at it.”