Amazon's AWS division today unveiled a new AI and speech-recogition tool intended to help doctors enter patient visit notes into their systems.

For now, AWS HealthScribe is only available as a preview in Northern Virginia (home of Amazon HQ2). But it promises to generate transcripts with "word-level timestamps" of patient visits, and automatically "identifies speaker roles, like patient and clinician, for each dialogue in the transcript," Amazon says.

This is something other AI tools also do, like Otter.ai for meetings, but AWS HealthScribe customizes the experience to the medical context.

"Clinicians often spend nearly twice as much time on administrative tasks instead of face-to-face interactions with patients," Amazon says, citing an American Hospital Association study. "This creates a struggle between providing compassionate care and maintaining accurate records."

HealthScribe generates summarized clinical notes with key sections like chief complaint, history of present illness, assessment, and treatment plan. Doctors can review the summary, make edits, and finalize the notes for clinical documentation.

The service is HIPAA-compliant and "does not retain inbound audio or output text, nor does it use your data to train AI models," Amazon says. "[Doctors] have full control to determine whether to store transcriptions in local environments or self-managed cloud storage."

Amazon is offering doctors a free preview of the tool—up to 300 minutes for two months. But after that, it requires payment by minute of audio analyzed, up to thousands of dollars per month depending on patient volume. An example on the pricing sheet says that an office with 1,000 patient visits per month, at 15 minutes each, will cost approximately $1,500.

It's the latest foray into the medical field for Amazon, which has a checkered past in the field. In 2020, it debuted Amazon Pharmacy, which is still up and running, while in August 2022 it shuttered the Amazon Care Telehealth business.

Generative AI feels like a natural next step for medicine technology, so it's no surprise Amazon is among the companies exploring it. Microsoft launched its DAX Express technology in March, which uses OpenAI's GPT-4 model to "reduce administrative burden and empower clinicians to spend more time taking care of patients and less time on paperwork." And Google is reportedly testing med-PaLM2, the medical version of the large language model behind Bard, in hospitals.

“It is clear that generative AI has the power to transform the health care and life sciences industry in many ways,” says Swami Sivasubramanian, AWS VP of database, analytics and machine learning services, as reported by CNBC.