Google expands open source artificial intelligence tools for healthcare imaging and transcription

Google has released MedGemma 1.5 and MedASR, two open-source medical artificial intelligence models designed for imaging workflows and speech-to-text dictation, as investment firms highlight accelerating artificial intelligence adoption in healthcare.

Alphabet has introduced two new open-source medical artificial intelligence models as part of its push into healthcare. The company launched MedGemma 1.5, described as a new generation of open-source healthcare artificial intelligence model focused on enhancing support for medical imaging capabilities, and released MedASR, a brand-new open-source automatic speech recognition model that is optimized for medical dictation. According to the report, these launches arrive amid growing interest from investors and technology companies in using artificial intelligence to improve clinical workflows and support commercial opportunities in healthcare.

The article traces MedGemma 1.5 back to last year, when Google used its Artificial Intelligence Developer Foundation project to debut the MedGemma open-source healthcare generative artificial intelligence model set, which was intended to give developers a starting point for evaluation and customization across different healthcare scenarios, and the model set has been downloaded millions of times. The latest update extends those models to a wider range of imaging applications, including high-dimensional medical imaging such as computed tomography (CT), magnetic resonance imaging (MRI), and histopathology, longitudinal medical imaging focused on retrospective time-series analysis of chest X-rays, anatomical localization that identifies anatomical structures in chest X-rays, and medical document understanding that extracts structured data from laboratory reports. MedASR’s initial version is described as allowing developers to convert medical voice input into text while integrating with MedGemma to perform advanced reasoning tasks.

Google cited data from Menlo Ventures to argue that healthcare is emerging as a leading adopter of artificial intelligence, with the healthcare industry deploying artificial intelligence at a rate 2.2 times faster than other industries. In this breakdown of adoption, health systems lead with 27%, ahead of outpatient facilities at 18%, while payers are reported at 14%. The article notes that large artificial intelligence companies and startups are moving aggressively into the sector, mentioning OpenAI’s launch of ChatGPTHealth, Anthropic’s new healthcare and life sciences services for its Claude artificial intelligence platform, and Baichuan Intelligence’s open-sourcing of Baichuan-M3, which achieved a comprehensive score of 65.1 in HealthBench and was described as ranking first globally. Brokerage commentary from CITIC Securities and Hua Fu Securities argues that the logic of artificial intelligence healthcare changed fundamentally in 2026 as paying entities became clearer and stronger, that artificial intelligence healthcare has fallen below the lowest point of the January uptrend, and that policy support, fundamental improvement, and product launches such as “Ant Alipay’s Afu” signal genuine demand and accelerating commercialization, with specific listed healthcare companies highlighted as recommended stocks to watch.

56

Impact Score

Pacific Northwest national lab unveils generative Artificial Intelligence cyber defense tool

Scientists at Pacific Northwest National Laboratory have developed a generative Artificial Intelligence system called Aloha that can rapidly simulate and adapt cyberattacks, offering defenders a faster way to reconstruct and test intrusions. The project highlights how U.S. national labs are turning years of classified Artificial Intelligence cyber research into practical defensive tools.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.