Generative AI is reshaping how people with speech impairments communicate, enabling the return or personality rhythm, and emotional inflection often lost with traditional total glossectomy and laryngectomy.
AI companies are building lifelike voice models from short recordings — sometimes just a voicemail — though 30 minutes of audio remains ideal, according to NPR. One of these tools, Whisper, allows users to input text and broadcast speech in their own AI-cloned voice through portable devices.
As more of these apps emerge and more researchers call for clinical trials, experts hope voice replication will gain recognition as a vital part of identity, potentially even insurance coverage.
Preserving a voice before losing it
When doctors told 51-year-old architect Sonya Sotinsky that the only way to save her from advanced oral cancer was to remove her tongue and voice box, she didn’t just prepare for the life-changing surgery, she recorded messages she hoped would help her stay connected after losing her ability to speak.
She spent her final five weeks before surgery capturing the everyday and the emotional: “Happy birthday,” “I’m proud of you,” and “I’ll be right with you” for her family and clients. She read children’s books in hopes of one day being part of her future grandchildren’s bedtime rituals.
Then, in a deliberate expression of personality, she also recorded long strings of expletives. If voice is a marker of identity, then sarcasm was going to survive the operation too.
“When you can’t use your voice, it is very, very frustrating,” she told NPR. “Other people project what they think your personality is. I have silently screamed and screamed at there being no scream.”
AI, emotion, and a natural-sounding comeback
After her total glossectomy and laryngectomy in 2022, early voice tech options which proved mechanical, flat, and barely intelligible, left her socially and emotionally isolated. But in 2024, Sotinsky discovered generative AI capable of recreating her vocal fingerprint: tone, inflection, even her New Jersey accent.
Using several hours of her personal recordings — especially those children’s book sessions — AI company Whisper developed a voice model that now lives in a mobile app. Sontinsky types, and the app speaks for her. Through portable speakers, her voice re-entered the room.
“Eloise saved my voice,” she said.
“When we heard her AI voice, we all cried,” added her daughter Ela Fuentevilla. “It’s crazy similar.”
What’s lost without a voice, what technology can restore
The loss of speech affects more than just one’s ability to communicate. It impacts thought processing, emotional expression, and social dignity. Patients like Sotinsky often face increased emotional distress, job loss, and social withdrawal.
“Communication is not only how we express ourselves and relate to other people, but also how we make sense of the world,” Dr. Sue Yom, a radiation oncologist at UC San Francisco, told NPR.
And yet, despite its obvious benefits, Sotinsky’s insurer denied reimbursement for her $3,000 AI voice technology. She now pays a $99 monthly fee out of pocket. “Apparently, having a voice is not considered a medical necessity,” she quipped.
AI as medical empowerment
AI didn’t just restore Sotinsky’s identity, it helped save her life in a different way. When her cancer returned, she used her AI voice to fully participate in treatment planning. “Being able to dialogue with my care team in a more seamless way is vital,” she said.
In a time when AI is often either hailed as a miracle or flagged as an ethical menace, Sotinsky’s story is a reminder: the most meaningful applications don’t replace humans, they help people remain more fully themselves.
As she put it… “Sarcasm is part of my love language.”
Healthcare organizations of all types are adopting generative AI in hopes of optimizing patient diagnoses, improving doctor-patient relationships, and providing administrative and clerical support in clinical settings. Read more about the changes and challenges GenAI poses in the healthcare space.
The post AI Tech Gives Cancer Patient Her Voice Back — With Sass Intact appeared first on eWEEK.
No Responses