The prize recognizes Stavisky's development of an AI-powered speech neuroprosthesis that turns brain activity into spoken words — giving people who have lost the ability to speak a way back into conversation.
For someone who can no longer speak, the loss is not only physical. It is the loss of the fastest, most human way to connect. Stavisky's work is aimed squarely at that loss. Through the system he and his team built, a participant living with amyotrophic lateral sclerosis (ALS) who could no longer speak intelligibly can now generate fluent sentences from his own brain signals — first as text on a screen, then as synthetic speech modeled on his voice from before the disease. In daily use over two years, the participant has spoken over 2.7 million words — and has resumed working full-time in climate advocacy, videoconferencing and writing to colleagues using only his brain signals.
The achievement sits at the center of a conviction the Chen Institute has held since its founding a decade ago: follow curiosity into the hardest questions. That pursuit is what advances frontier science and carries its breakthroughs to the people who stand to benefit from them.
"Ten years ago, Tianqiao Chen and I founded the Chen Institute around a single question, not an answer: how does the brain give rise to intelligence? Where that question would lead was impossible to know, but that is the nature of real discovery. Dr. Stavisky set out to solve a problem many in his field considered unsolvable, and in doing so gave people back something profoundly human: their own voice. A decade on as we celebrate our 10th year milestone, the conviction behind the Institute has not changed: the questions worth asking are the ones without easy answers and are inherently discoverative in nature; the scientists worth backing are the ones willing to sit with the unknown long enough to change it. The Institute remains committed to championing the researchers redefining what science can achieve."
— Chrissy Luo, Co-founder, Tianqiao and Chrissy Chen Institute
HOW IT WORKS
Four small arrays — 256 microelectrodes in total — are implanted into the brain region that coordinates speech. As the participant tries to speak, the device reads the firing patterns of hundreds of neurons. One AI model translates those patterns into phonemes, the basic sound units of language. A second model, built on the same large-language-modeling approach behind today's AI assistants, assembles those phonemes into words and sentences. A newer "brain-to-voice" method skips text entirely and reconstructs the sound of the person's own voice as they try to speak — fast enough that the conversation feels less like sending a text and more like a live phone call.
THE NUMBERS
In a study published in the New England Journal of Medicine (2024), the system decoded intended speech with up to 97.5% word accuracy across a 125,000-word vocabulary — accuracy comparable to, or better than, consumer speech-recognition tools. In a follow-up study published in Nature (June 2025), the team achieved real-time voice synthesis with a delay of about 30 milliseconds — close to the natural delay between speaking and hearing your own voice. The synthetic voice can carry intonation, ask questions, and even sing. In a long-term home-use study published in Nature Medicine (June 2026), the participant successfully sustained full-time employment while using the system to communicate.
Stavisky's focus on speech came from watching what patients valued most. Earlier in his career, working on brain-computer interfaces for movement, he noticed a pattern: restoring the ability to move a cursor or a robotic arm mattered, but restoring communication was usually more urgent. "Communication was almost always the number one priority," he says. That realization, paired with advances in machine learning and brain-recording technology, led him to change course mid-career toward what was then considered one of the hardest problems in the field.
"Stavisky developed an AI-based speech neuroprosthesis with immediate and transformative practical impact. It restored communication for a paralyzed patient with ALS with over 99% word accuracy, enabling the patient to express over 2.7 million words over two years using only brain signals. His team achieved real-time voice synthesis, allowing the patient to modulate intonation and even sing."
— Yury V. Suleymanov, senior editor at Science
The field is now moving quickly. Companies are entering clinical trials for speech brain-computer interfaces, and academic labs are testing whether similar approaches could help people with stroke-related aphasia, cerebral palsy, and other conditions that take away speech. Stavisky's long-term goal is a surrogate voice so natural that, on a phone call, you could not tell it was synthetic — delivered through devices that are smaller, fully implanted, and ready to move from research prototype to everyday clinical tool.
That trajectory — from an unsolvable problem to a working voice — is the kind of long-horizon, discovery-driven science this prize exists to recognize.
2026 AI Prize Finalists
The 2026 Prize also recognized two Finalists whose work is reshaping what AI can do across the sciences — from mental health to materials discovery:
Dr. Nicholas C. Jacobson — Dartmouth College
Associate Professor of Biomedical Data Science and Psychiatry, Geisel School of Medicine
Dr. Jacobson was recognized for developing Therabot, a generative AI system that delivers personalized, evidence-based psychotherapy for anxiety and depression. Built over 6.5 years with more than 100,000 human hours of clinical oversight, Therabot completed the first randomized controlled trial of AI-delivered psychotherapy — achieving symptom reductions with effect sizes approaching the best evidence-based treatments across all three clinical domains tested. His work addresses a global treatment gap where the median number of mental health providers worldwide is only 13 per 100,000 people. Therabot points toward a future where geography and cost are no longer barriers to quality mental healthcare.
Dr. Zach (Zhiling) Zheng — Washington University in St. Louis
Assistant Professor of Chemistry
Dr. Zheng was recognized for building a general-purpose AI laboratory system driven by autonomous agents that unifies literature mining, generative molecular design, and robotic experimentation in a single closed loop. Applied to metal-organic frameworks (MOFs) for water harvesting, his generative design agents proposed 10 new materials — the "long-arm" MOF family (LAMOF-1 through LAMOF-10). In a closed-loop robotic demonstration on two of these materials, his AI agents identified high-crystallinity conditions within 2 weeks. They mine the scientific literature, propose novel molecular structures, and autonomously navigate and interpret experiments across a reaction space of more than 6,000,000 possible combinations, compressing the timeline from design to discovery. His platform offers a blueprint for how AI can accelerate materials science across clean energy, water, and health applications.
Together, this year's winner and finalists represent a generation of scientists using AI not as an end in itself, but as a means to reach problems that once seemed out of reach — restoring lost voices, expanding access to mental healthcare, and accelerating the discovery of materials the world needs.
ABOUT THE CHEN INSTITUTE & SCIENCE PRIZE FOR AI ACCELERATED RESEARCH
The Chen Institute & Science Prize for AI Accelerated Research celebrates outstanding early-career researchers who are advancing interdisciplinary breakthroughs at the intersection of artificial intelligence and natural sciences. By recognizing bold ideas and elevating innovative work across diverse scientific domains, the Prize highlights the next generation of thinkers driving discovery in the AI era. For more information, visit cheninstitute.org/aiprize.
ABOUT THE TIANQIAO AND CHRISSY CHEN INSTITUTE
The Tianqiao and Chrissy Chen Institute is a global, multidisciplinary organization dedicated to advancing fundamental brain science, accelerating scientific discovery through artificial intelligence, and improving human well-being. Founded with the mission to understand how the brain perceives, learns, and interacts with the world, the Institute brings together leading researchers, physician scientists, engineers, and innovators across premier institutions worldwide. For more information, visit cheninstitute.org.
SOURCE Tianqiao & Chrissy Chen Institute
Share this article