Questions Arise Over Efficacy of AI Healthcare Tools

AI-powered health management tools are rapidly proliferating in the market, yet questions are emerging about their actual effectiveness and safety. In the high-risk healthcare sector, critiques are mounting that relying solely on corporate self-assessments or publicizing evaluation results without external expert review is untrustworthy. Concerns are also raised that companies might miss potential blind spots where they could benefit from the research community. Experts emphasize the need for more rigorous and systematic evaluations by independent professionals before these AI healthcare tools are widely launched.

AI tools in the form of specialized chatbots are gaining attention as an alternative for providing health information, particularly for those with limited access to existing medical systems. Some research findings suggest that certain LLMs (Large Language Models) can offer safe and useful health advice. AI developers are releasing healthcare products, believing that their large language models have reached a level capable of providing medical guidance.

Dominic King, Vice President of Microsoft's AI health division, stated that advancements in AI capabilities are the core reason for the formation of his health team and the existence of Copilot for Health. He noted improvements in generative AI's ability to answer and respond to health-related questions, adding that demand is another key factor. Just before the launch of Copilot for Health, Microsoft released reports and blog posts detailing how people were using Copilot for health advice, indicating that Copilot received 50 million health-related queries daily, and health was the most popular conversation topic on the Copilot mobile app.

Other AI companies have also launched similar products, noting this trend.