By Bob Katzen

House 12-145, rejected an amendment to a bill that would prohibit the distribution of deceptive artificial intelligence (AI) communications within 90 days of an election, including audio or visual media which depicts a candidate with intent to injure their reputation or deceive a voter into voting for or against them.

The amendment would narrow the language to allow the use of AI for translation and transcription.

Amendment sponsor Rep. John Gaskey (R-Carver) said the bill travels dangerously close to the edge of First Amendment violations, especially when we’re talking about challengers and grassroots organizations that have smaller budgets than establishment and incumbent positions. He noted that AI is now being used as a low-cost tool for these organizations to be able to generate transcriptions and translations to get their information out to more people, and in multiple languages.

Rep. Tricia Farley-Bouvier (D-Pittsfield) opposed the amendment and said that the bill does nothing to affect free speech. She noted that the measure refers to misinformation, not just of a candidate, but of the underlying infrastructure that we need to trust to have elections. Like where the polling place is and what the times are. It talks about actual malice and intent. It’s not a mistake somebody makes, it’s about intent.

(A “Yes” vote is for the amendment. A “No” vote is against it.)

Rep. Christine Barber No Rep. Mike Connolly No Rep. Paul Donato No Rep. Erika Uyterhoeven No