Canary Speech

Canary Speech recently partnered with NeuroLexIQ to explore how our vocal analysis can support earlier awareness of potential mild traumatic brain injuries. By integrating our technology into NeuroLexIQ's intake workflow, subtle speech patterns can help flag when further evaluation may be warranted. This is one of many contexts where objective, voice-based signals can complement existing assessment methods, and a natural extension of what we've spent the past decade building: turning voice into a clinically meaningful signal. We're excited to watch this partnership grow! Learn more about our work here https://lnkd.in/g82YZz6w and at canaryspeech.com.

Why does Parkinson's affect more men than women? A new study covered by Medical News Today examined five brain regions and found that while Parkinson's triggers shared "stress responses" across everyone's brain cells, there are meaningful differences between men and women at the cellular level — particularly in how the brain's support cells manage energy and protect nerve connections. Our Chief Medical Officer, Dr. Kang Hsu, MD, weighed in: "What stands out most to me is that women with Parkinson's showed stronger blood-based gene-switch changes than men, even though men get Parkinson's more often." Findings like these underscore why Parkinson's screening and monitoring shouldn't treat every patient as biologically identical. It's exactly the kind of nuance that motivates our work. At Canary Speech, our Parkinson's model uses vocal biomarkers, how sound is produced and not the words being said, to screen for Parkinson's from natural, conversational speech. In research with Beth Israel Deaconess Medical Center and UMass Chan Medical School, our approach detected Parkinson's with 97% accuracy, non-invasively and at scale. As the science reveals more about how Parkinson's differs between individuals, objective, accessible tools will be essential to catching it earlier for everyone. https://lnkd.in/gEHycisd #Parkinsons #VocalBiomarkers #DigitalHealth #AIforGood

The hardest conditions to catch are the ones that hide in plain conversation. Anxiety, depression, and early cognitive change rarely announce themselves. They surface in subtle shifts in how a person speaks. Canary Speech listens for those shifts. Paired with eNcounter®, it adds objective behavioral health screening to visits your providers are already having. Schedule a demo: https://lnkd.in/gF4QUZEQ #GlobalMed #Telehealth #DigitalHealth #PatientCare #ConnectedCare

Identifying at-risk patients earlier is one of healthcare’s biggest opportunities, and one of its most difficult challenges. In our latest blog, we explore how healthcare organizations can move from reactive intervention to proactive care by reducing administrative burden, embedding actionable AI insights directly into clinical workflows, and using data synthesis to identify risk earlier. When clinicians have the right information at the right time, they’re better equipped to intervene in the moment and improve patient outcomes. Read the full blog: https://lnkd.in/gbXERR36 #CanarySpeech #AIforGood #ClinicalWorkflows

Happy Fourth of July! Wishing our US customers, partners, colleagues, and communities a safe, relaxing, and joyful Independence Day. We hope you have the opportunity to celebrate with family, friends, and those who matter most. Have a wonderful holiday!

We are excited to share that Canary Speech's latest research has been accepted to AIME 2026, the International Conference on Artificial Intelligence in Medicine, taking place July 7–10 at the University of Ottawa. Our VP of Research, Namhee Kwon, will be presenting our peer-reviewed paper: "Automated Behavioral Health Assessment Using Vocal Biomarkers." This study demonstrates that a conversational speech sample can screen for depression and anxiety with clinically meaningful accuracy. * 0.80 specificity 0.68 sensitivity on an independent in-clinic dataset * Validated against PHQ-8 and GAD-7 clinical instruments * Works from a simple "How are you?" prompt, no specialized tasks required Depression and anxiety affect over 690 million people worldwide, yet early identification remains one of healthcare's biggest challenges. This research shows that the voice itself (not what you say, but how you say it) can serve as a non-invasive, language-independent screening signal, usable in-clinic or remotely. This is part of Canary Speech's ongoing commitment to advancing the science of vocal biomarkers. From clinical documentation to behavioral health, our research team continues to publish at the frontier of voice AI in medicine. We'll see you in Ottawa! Read the paper here: https://lnkd.in/gQMtcy_R  #AIME2026 #CanarySpeech #AIforGood

Big news: Canary Ambient™ is now live on the Zoom App Marketplace. For the millions of clinicians already using Zoom to deliver virtual care, that means real-time vocal biomarker insights are now available directly inside the telehealth visits they're already running. Canary Ambient™ analyzes more than 2,500 speech features in real time, surfacing objective signals for Depression, Anxiety, Mild Cognitive Impairment (MCI), and Alzheimer’s Disease, often before traditional clinical measures or noticeable symptoms appear. It helps clinicians move past "I'm fine" to uncover what's actually going on. Telehealth has become a primary access point for behavioral and cognitive care. But virtual visits have historically missed the subtle vocal and linguistic cues that matter most. This integration turns routine conversations into early indicators of risk and opportunities for recovery. Read the full announcement here: https://lnkd.in/gXi8HJjm Canary on Zoom Marketplace: https://lnkd.in/gUhjpSUD #CanarySpeech #ZoomHealthcare #Telehealth #AIforGood

Canary Speech's CTO, Nate Blaylock, and VP of Engineering, David Brown, will be presenting at the Zoom Developer Summit 2026. This virtual event built for developers features only a handful of partner companies, and we are excited and honored to share our insights on real time media streams. The session, "Turning Zoom RTMS into Voice Based Health Insights with Canary Speech," will cover how we use Zoom's technology to surface clinically meaningful data from voice within the Zoom environment. Real-time communication is the foundation for intelligent health applications, and we are looking forward to sharing our voice! Register here: https://lnkd.in/gNgX-k4y #ZoomDeveloperSummit #Zoom #CanarySpeech

Canary Speech has published its second model card on the Coalition for Health AI (CHAI) Registry. This card documents our Cognitive Health model, a clinical decision support tool that uses voice biomarkers to help clinicians identify Mild Cognitive Impairment and Alzheimer's Disease in patients 60 and older. The CHAI registry was developed with Avanade and is supported by leading health systems including Cleveland Clinic, Kaiser Permanente, and Providence. It provides a standardized framework for evaluating health AI, covering intended use, patient population, performance, fairness considerations, and known limitations. Model cards serve a specific purpose in responsible AI deployment: they create a documented, accessible record of what a model does, how it was developed, and where clinical judgment remains essential. For health systems evaluating AI tools, that documentation is foundational to informed purchase and safe implementation. Publishing a second card reflects our commitment to doing this across all the clinical domains our technology operates. View both model cards on our registry page: https://lnkd.in/gn_Q3-UQ