The Jerusalem PostTech
Israeli Study Tests AI Chatbots on Medical Test Recommendations
Summarized from The Jerusalem Post by baba
Center-rightNeutral tone
An Israeli study by Ben-Gurion University and Clalit Health Services found that leading AI chatbots can identify unnecessary medical tests and treatments with high accuracy. The LLMs made errors in only 0.6% of cases where care was needed and 5.4% where it was not. However, AI struggled with ambiguous cases, mirroring human medical uncertainty. Researchers stressed that doctors remain responsible for patient care, with AI serving as a tool.
The story in 5 lines · by baba
- Leading AI chatbots accurately identified unnecessary medical tests and treatments in an Israeli study.
- AI models erred in only 0.6% of cases where care was needed and 5.4% where it was not.
- AI struggled with ambiguous clinical scenarios, such as EKGs for palpitations and ultrasounds for thyroid nodules.
- Researchers noted AI's tendency to agree with patients, unlike human physicians in delicate situations.
- The study concluded that doctors remain responsible for patient care, with AI acting as a tool.
The Jerusalem Post publishes in English. The key points are summarized above. Read The Jerusalem Post’s full report from the link at the top of the page.