The French startup Bioptimus is releasing an artificial intelligence model trained off of hundreds of millions of images that, it said, will aid in the research and diagnoses of diseases. The model, called H-optimus-0, is capable of performing complex tasks including identifying cancerous cells and detecting genetic abnormalities in tumors, the Paris-based company said in a statement. Bioptimus described the system as the largest model for pathology that is open source and widely accessible. The startup is among the many firms trying to leverage AI for medical breakthroughs. Google’s DeepMind recently released a new version of AlphaFold, a landmark AI system developed for predicting protein structures. K Health, an American startup behind a chatbot that speaks to patients before they see physicians, raised $50 million in a transaction that values the company at about $900 million. Bioptimus Chief Technology Officer Rodolphe Jenatton said the company opted for an open-source model to bring about transparency in the space and accelerate advances. The company said it hopes further developments will allow the model to work at different levels of human biology. (The current model focuses on tissues only.) Widespread concerns about the use of AI in diagnosing medical conditions remain. A 2023 Pew Research Center survey found that 60% of patients would feel uncomfortable with their doctors relying on AI for their health care.
Full report : French Startup Bioptimus Releases AI Model for Disease Diagnosis.