Use of artificial intelligence in campaign ads draws scrutiny

Artificial intelligence has been used in ads in statewide, legislative, and congressional races in Pennsylvania, including digital representations of candidates saying words they did not actually say.

Political ads have long featured attacks on candidates using look-alike actors, digital animation, and images edited or manipulated with programs like Photoshop.

But this year's campaign ads in Pennsylvania have featured likenesses of candidates generated with artificial intelligence---sometimes saying things they've never said.

The use of that technology in ads for statewide, congressional, and legislative races has sparked discussion---and scrutiny or criticism among some--- about the appropriateness of using generative A.I. in political ads and content.

"We're always deciding where to draw the line---that's part of the debate in a healthy democracy. But I do think that putting words into someone else's mouth is not necessarily protected speech," said Berwood Yost, director of Franklin and Marshall College's Center for Opinion Research.

Charles Palmer, executive director of Harrisburg University's Center for Advanced Entertainment & Learning Technologies, said that the use of generative A.I. by campaigns remains in the "Wild West" phase but that the "barn doors are open."