“My new – old – AI voice,” he said in a recent video introducing the voice to his constituents.
Wexton made headlines this year for using a robotic-voiced speech application to deliver remarks on the House floor. It was a widely appreciated display of resilient spirit, but App It didn't sound like it.
This week, Wexton flashed her new, more natural voice as she stood to address the House Appropriations Committee, occasionally touching the walker she's been using since being diagnosed with a progressive Parkinson's-like disease. called supranuclear palsy (PSP).
“For those of you who heard me speak before the PSP took away my voice, you may think your ears are deceiving you,” he told the committee via an app running on his iPad. Giving.” “I assure you they're not. I'm using a new AI model of my voice today—I know, it's cool.
The cadence, the tone, the timbre all sounded remarkably like Wexton, who spent five years in the Virginia Senate before being elected to Congress in 2018. He is not seeking re-election this fall. Because of his health condition.
For his crew, the moment was more chilling.
“It was a big deal. It was something we really didn't expect to be able to hear again,” Justin McCartney, Wexton's director of communications, said Friday.
Wexton, 56, has not lost any of his wits and intelligence, McCartney said, but he has difficulty communicating verbally because of a neurological disorder. For those who have been close to Wexton over the years, A.I He said the sound brought back something that many people didn't realize how much they missed.
“It's never going to be 'me,' but it's more than I or anyone around me thought we'd hear again,” Wexton said Friday in a written response to questions from The Washington Post. After turning down speaking engagements and public appearances because she could no longer trust her voice, Wexton said, technology has restored her ability to “continue to do my best in this job.” Is.”
The first sample he heard of his AI voice was a snippet of Hamlet's “To be or not to be” dialogue. “My husband had a big smile on his face,” she said. “I've never seen it so broad and so real [happy] In a very long time,” he said, adding that he enjoyed seeing the reactions of friends and colleagues to it.
“It was an incredibly moving moment to hear my friend's voice again,” Rep. Abigail Spinberger (D-Va.) said via email Friday. “And while I know this new voice is made possible by a computer, I know the words of wisdom — and smart but cutting jokes — we'll hear from Jennifer and Jennifer alone.”
Late last year, when Wexton wanted to address the House in support of a bill aimed at ending Parkinson's disease, he wrote remarks and had them delivered to him by Rep. Jennifer McClellan (D-Va.). What did It was a humbling moment, McClellan said in an interview with The Post on Friday.
“For an independent person who made a career out of using his voice on behalf of other people — first as a prosecutor, then as a legislator — relying on someone else to speak on your behalf is very difficult. It was difficult.” McClellan said. “I think the AI software gave it back some of its power in a way that was very dynamic.”
In his comments to The Post, Wexton acknowledged the pitfalls of AI, saying that “it's scary to think about the evil things that someone with bad intentions could do with this technology.” He has limited who on his team has access to the tool, “because using my voice to say something I don't want to say can cause real problems.”
Experts have been quick to sound the alarm about technology's potential to disrupt politics. A report from the Wilson Center in January warned that with national elections scheduled in countries around the world this year, “the threat of blurring the walls of reality, as one analyst put it, created by AI And by using a lot more. Traditional deepfake productions are disturbingly high.”
The Brennan Center for Justice at New York University School of Law noted late last year that generative AI is already being used in political ads and mailers in the United States and is “poised to redefine the modern campaign.” The Center's report said Congress has not yet addressed the issue.
McClellan said Wexton's adoption of AI underscores the need to bring laws and regulations up to date.
A video of the Wexton voice model using the text-to-speech app in May caught the attention of New York-based AI voice startup ElevenLabs. The company contacted the congresswoman's staff, and after giving her the green light, the staff spent a few weeks compiling more than an hour of audio clips of Wexton speaking before the illness affected her voice.
It only took the company a few days to create a digital version of Wexton's voice. An ElevenLabs spokesperson said that artificial intelligence makes it possible to not only mimic a voice, but also change the intonation so that it sounds natural rather than robotic.
“Our model is able to understand the relationships between words and adjust delivery based on context … to produce lifelike, human-sounding speech,” ElevenLabs spokesman Sam Sklar said via email.
Wexton's office pays ElevenLabs a small subscription fee for the service, McCartney said.
ElevenLabs, launched in 2022, has approached other public figures with similar offers, such as former Spirit Airlines CEO Ben Baldanza, who There's Amyotrophic Lateral Sclerosis, and Lori Cohen, a lawyer at Greenberg Trauerig, who uses an AI version of her voice to argue cases in the courtroom.
“It's clear that work remains to be done to adequately protect against the potential threats this poses,” Wexton said in his video. “But it can also provide new, unimaginable and life-changing opportunities for Americans with disabilities.”