UC Davis Researchers Develop AI-Augmented Video Tool to Improve Depression Screening
Researchers at the University of California, Davis, have developed a new artificial intelligence (AI)–powered video assessment tool, Async, designed to enhance the accuracy and accessibility of remote mental health and medical screenings. The platform integrates machine learning, multi‑language patient interviews by AI video agents, ambient AI, and voice and movement recognition technologies. The goal is to enable early identification of clinical depression and other behavioral or medical conditions.
By merging AI‑driven analysis with clinician expertise, Async enhances early diagnostic insight with the goal of expanding access for patients who face barriers to in‑person healthcare.
During the COVID‑19 pandemic, restrictions on in‑person medical visits accelerated the use of telepsychiatry for cost and access reasons. Since then, online screening and consultation have become increasingly acceptable to clinicians, patients and insurers.
While machine learning has shown promise in curating patient histories and predicting health risks, many early remote‑assessment systems have delivered inconsistent results or struggled to adapt to the linguistic and cultural needs of diverse patient populations. To address these limitations and enhance the quality of remote clinical engagement, researchers at UC Davis carried out over a decade of clinical trials resulting in the development of the Async platform.
Peter Yellowlees, distinguished emeritus professor of psychiatry at UC Davis Health, and his team launched this Async platform through AsyncHealth, a spin-off corporation from UC Davis, leading to UC Davis granting an exclusive license to the company for the patent rights in the technology.
Over $7 million of federal funding from the Agency for Healthcare Research and Quality for three clinical trials played an important role in the early stages of technology development. The team worked closely with the Technology Transfer Office within the Office of Research, which oversees patent protection and licensing, and completed the prestigious NSF I-Corps program to focus on product market-fit, as well as taking part in the UC Berkeley Skydeck accelerator program. “Developing Async over years of research and technical development has been a great learning experience for us,” said Yellowlees.
How does the platform work
As a healthcare platform, Async is designed to ease pressure on overloaded medical systems. Instead of relying on traditional in‑person visits for intake, follow‑ups, and chronic‑disease check‑ins, the platform lets patients share information and complete routine monitoring on their own time through secure digital tools. By separating basic data‑gathering from live clinician time, the platform acts like a virtual assistant that supports doctors around the clock. The company states that healthcare practices can use it to expand access, improve care quality, and free up clinicians’ schedules—without hiring more staff or extending office hours.
One of the features of this video interviewing platform is that it lets patients complete secure clinical assessments without scheduling an appointment. Patients record short video responses to a set of tailored, pre‑recorded questions delivered by lifelike AI video agents. Their responses are automatically transcribed, summarized and turned into an initial medical note by an AI scribe. Clinicians then review this information on a dashboard to triage patients to the right level of care or see them with most of the history and documentation already drafted.
The system, researchers note, is designed to save clinicians time while remaining low‑cost, distributed, and easily scalable. The team explains that Async integrates directly into existing billing workflows, allowing clinics to continue receiving reimbursement exactly as they do today.
Media contact: Neelanjana Gautam, UC Davis Office of Research, [email protected]
