Global AI Healthcare Projects Expand

This week, the European Union highlighted its investment in AI-driven healthcare projects to improve maternal and newborn care. Programs like SHAIPED focus on faster approval of AI medical tools, while others train workers in digital skills for hospitals. In the U.S., the FDA released draft guidelines for AI-enabled medical devices, emphasizing the need for companies to monitor and update their tools over time. This includes software that analyzes X-rays or manages patient data.

AI in Emergency Care

Medtronic announced a partnership to develop AI systems for stroke care. Their technology scans brain images to detect blood clots, helping doctors decide on treatments like surgery within minutes. Early tests show this could reduce brain damage by speeding up decisions. Meanwhile, Mount Sinai researchers found that AI models sometimes recommend costly tests (like MRI scans) more often for high-income patients, while low-income patients get fewer tests. The study urges hospitals to check AI tools for fairness before using them.

Mental Health and Precision Medicine

Startups are using AI to create virtual therapists that chat with users about anxiety or depression. These tools analyze words and voice tones to suggest coping strategies. For serious illnesses like cancer, AI helps match treatments to a patient’s genes. A VC expert noted this could cut trial-and-error in drugs, especially for rare diseases. In Japan, GE HealthCare’s new purchase aims to blend radiopharmaceuticals with AI for better cancer imaging.

Policy and Ethics Debates

The U.S. Department of Health wants to replace lab animals with AI models for drug testing, which could save time and reduce animal use. However, critics worry digital simulations might miss some side effects. At a health law conference, experts argued for strict rules to prevent AI bias and leaks of patient data. Sessions covered topics like “who’s responsible if an AI makes a mistake?” and how to audit algorithms.

Research and Global Collaboration

A Icahn School of Medicine study tested nine AI language models on 1,000 fake patient cases. The models changed triage recommendations based on race or income, highlighting the need for better training data. In another project, Cognizant reported that 70% of companies feel they’re too slow to adopt AI, but those who do see major gains in tasks like scheduling or billing.

Lastly, Boston Scientific closed a deal to buy Bolt Medical, which uses AI to guide heart procedures. This follows a trend of big firms snapping up AI health startups to stay competitive.