AI is everywhere these days. Can it help catch nomination petition shenanigans?

Candidates seeking spots on the May 16 primary ballot are now more than a week into the arduous three-week season of asking voters to sign nomination petitions.

Then the real fun begins after the March 7 submission deadline, as competitors parse petitions, looking for ways to get opponents booted from the ballot.

Joe Driscoll thinks he has found a way to make it a little easier, with the help of artificial intelligence, also known as AI.

Driscoll, a former deputy city commissioner, developed a system of off-the-shelf AI products to compare petitions to voter registration records. That could help find challengeable signatures — like a Republican who signed a Democrat’s petition or a suburban voter who signed for a city candidate.

Driscoll said that could free campaigns to spend more time looking for the real shenanigans in petition season, like “kitchen-table jobs,” when petition circulators skip the circulation and use voter lists to forge names and signatures.

Kevin Greenberg, an elections lawyer who conducted a continuing legal education course last month on the use of AI, is cautiously curious about Driscoll’s approach. He thinks the system could catch signatures from voters who signed more than one petition for a single office, like mayor, when they are only allowed to back one candidate.

“There are things that people who know how to do this do that this won’t replace,” Greenberg said. “But I’d love to see where it evolves to.”

Adam Bonin, another election lawyer, also sees potential as a tool for determining a candidate’s rate of success or failure in turning in legitimate petitions. But that won’t take humans out of the equation.

“You’re still going to want to do a careful line-by-line review before you go before a judge,” Bonin said.