When Jennifer Wexton stands up to speak on the House floor on Thursday, something she has done countless times before, the congresswoman will wield a voice she thought was gone forever.
After a rare neurological condition robbed her of her ability to speak clearly, Wexton has regained her voice with the help of a powerful artificial intelligence program that allows the Virginia Democrat to create a clone of her speaking voice using old recordings of speeches and performances she gave as a congresswoman. She will use the program to deliver what is believed to be the first speech on the House floor ever delivered in a voice cloned by artificial intelligence.
“It was a special moment that I never thought would happen. I cried tears of joy when I first heard it,” Wexton told The Associated Press in the first interview she has given since getting her new voice.
Wexton’s voice came from her iPad, backed by a rainbow floral case on her dining room table in Leesburg, Virginia. The congresswoman types her thoughts, uses a stylus to move the text around, presses play, and then the AI program transcribes that text into Wexton’s voice. It’s a lengthy process, so the AP gave Wexton a few questions before the interview to give the congresswoman time to type her answers.
Wexton was diagnosed in 2023 with progressive supranuclear palsy, an aggressive neurological condition that affects many aspects of life, including speech. Sitting across from a dresser covered in photos of highlights of her personal life — weddings, family trips, her children — the congresswoman called the diagnosis “cruel” for someone whose “entire professional life has been built around using my voice,” from Virginia prosecutor to senator to member of Congress.
“A politician who can’t speak in public will soon be a former politician. But this AI voice model has given me a new opportunity to make my voice heard and reminds listeners that I’m still me,” Wexton said.
The congresswoman, whose landslide victory in 2018 signaled the success Democrats would have that year, initially announced a diagnosis of Parkinson’s disease in April 2023 and struck an optimistic tone, telling her supporters they were “welcome to feel empathy for her” but not to “feel sorry for me.” Her tone in September 2023 was very different, describing her Parkinson’s diagnosis as “Parkinson’s on steroids” and saying she would not seek re-election in 2024.
“This new diagnosis is a tough one. There is no ‘getting better’ with PSP. I will continue with treatment options to manage my symptoms, but they do not work as well for my condition as they do for Parkinson’s,” she said at the time.
The diagnosis changed Wexton’s personal and professional life. The congresswoman no longer looks like she once did. Her posture has slumped, her movements have become less precise, her natural voice has become muffled—all effects of the disease. As it became harder for Wexton to use her voice, she turned to a traditional text-to-speech app that many people with speech disorders often use. The voice sounded more like a robot than a human, but Wexton used it to conduct interviews and deliver speeches.
ElevenLabs, a startup with one of the most widely used AI-driven voice cloning models, saw Wexton speaking using the older technology. They reached out to her office a few weeks ago, and Wexton’s aides provided the company with several recordings, mostly of speeches she had given as a member of Congress.
“Our technology gives people who have lost their voices the ability to speak as they once did, with the emotion and passion they feel, and we hoped to help Congresswoman do just that,” said Dustin Blank, the company’s Head of Partnerships. “She is an incredible public servant and a remarkable leader for people with disabilities.”
Wexton said she first used the cloned voice to speak with President Joe Biden in the Oval Office earlier this month as he signed the National Plan to End Parkinson’s Act, a bill Wexton called the “most consequential action we’ve taken in decades to combat Parkinson’s and related diseases like my PSP.” Days later, Wexton publicly debuted her cloned voice in a video, prompting an outpouring of support and thrusting the congresswoman into a debate over artificial intelligence.
“This is never a situation I envisioned myself in” and it’s “not the way I thought I would leave Congress,” she said. “I didn’t expect to be at the forefront of a debate about the future of AI.”
The use of AI-driven cloning to give Wexton her vote back is one of the positive uses of this technology. However, voice cloning has also been used in malicious ways, such as scamming people and pushing fake political messages. The most notable of these cases was when an AI-generated robocall impersonating President Joe Biden urged voters not to vote ahead of the New Hampshire primary. The call was quickly reported and had dire consequences for those behind it, but the incident raised serious questions about the future of this technology and the companies behind it.
Wexton, whose district is home to dozens of data centers that power AI, has similar questions. After debuting her voice clone, Wexton jokingly sent a few friends the same message: “AI isn’t all bad, but it’s mostly bad.”
Hany Farid, a professor and expert in digital forensics at the University of California, Berkeley, said Wexton’s example is an exception to the many malicious applications of voice cloning technology.
“I found it really moving… and I’m all for this application,” he said. “But I just want to emphasize that just because there are these really beautiful stories… that doesn’t mean we should ignore the really nasty things that are happening with these technologies.”
One way to ensure the technology is used for good, Farid said, is to have “better checks and balances” to ensure “people don’t do bad things with your products.” That includes content credentials that indicate how the audio was developed, storage of all audio created using the technology and know-your-customer rules that require companies that clone speech to know who is using their technology.
Wexton agrees that more guardrails are needed. Her team of advisers has taken precautions to ensure her likeness is protected, from limiting access to the voice to just three people to tightening the program’s security.
“It’s human and it’s empowering. It can also be dangerous,” she said of her new voice. “I still believe that the dangerous potential of AI technology needs to be better understood and that steps need to be taken to prevent the misuse of the technology, like deepfakes, from spreading, and part of that falls on lawmakers like us in Congress,” she added later.
In 2019, Wexton won bipartisan approval for an amendment directing the National Science Foundation to conduct research into public awareness surrounding deepfake videos generated by AI.
Wexton also said the technology isn’t perfect. Because the audio used comes from speeches and public events, it’s not great for casual conversation, often making everything sound like “one big proclamation.” Her two college-age sons, she said, don’t like it for that reason and, she joked, she doesn’t use it to “ask my husband to pass me the ketchup,” showing a sense of humor she’s known for on Capitol Hill.
“Ultimately, it will never be me. But it is more me than I could ever hope to hear again and for that I am so grateful and excited,” she said. “I plan on making the best of it.”
For doctors like Jori Fleisher, director of the Rush CurePSP Center of Care, that’s why this kind of technology can be life-changing for people diagnosed with this rare neurological condition.
Too often, PSP patients lose their voices and must rely on traditional speech-to-text programs to communicate, Fleisher said. But those programs use robotic voices that often sound nothing like the patients. Fleisher notes that people with “neurological diseases are already stigmatized,” so speaking in a voice that sounds like a computer “perpetuates the stigma” and often leads to withdrawal from relationships and “exacerbates the social isolation that can be such a big part of these conditions.”
“To know and have such a deep respect for Representative Wexton and then to hear her speak so beautifully in her own voice, using her own words through this technology, it gives me goosebumps right now,” she said, growing emotional. “It’s so encouraging.”
The key, Fleisher added, is to make this technology available to more people by encouraging patients in the early stages of PSP and other neurological disorders to “store enough of the sounds of your own voice that it can be used later” and having insurance companies cover such treatments. Wexton said she tried to do this through an Apple program late last year, but her voice was already too affected by the disease for their AI to work.
“This is an aspect of care and quality of life that can be dramatically improved and should be covered by a person’s insurance. If speech therapy is covered, as it should be, this is an extension of that,” she said. “This is … helping them be whole.”
Wexton’s new voice especially helps in more emotional moments, when hearing sentiment in her speech is considerably more powerful than a more robotic tone. When asked how Barbara Comstock, the Republican congresswoman to whom Wexton has become close since defeating her in 2018, had supported her since Wexton announced her diagnosis, the Democrat became emotional and said, “She’s been so gracious.”
“I was just happy for her,” Comstock said, recalling the first time she heard Wexton’s AI voice. “It’s amazing to hear her literally use her voice so that others can see the power of the technology. … It brings tears to my eyes just thinking about it.”
After defeating Comstock in 2018, Wexton’s future in Virginia politics looked bright, with many in the state speculating that she might seek higher office. Her diagnosis has robbed her of that future—her political career ends next year—but it has given Wexton new resolve.
“I want to be a voice, even an AI voice, for Americans who face accessibility issues and other disabilities, because too often people only see us because of that disability,” Wexton said. “I hope that by continuing to do the best job I can, whether that means using a walker or a wheelchair to get to the House floor to vote or delivering my speeches with an AI-rendered version of my voice, that it can help show that I am just as myself on the inside as I have always been.”
—
The Associated Press receives funding from the Omidyar Network to support its coverage of artificial intelligence and its impact on society. AP is solely responsible for all content. Find AP’s standards for working with philanthropies, a list of supporters and funded coverage areas at AP.org