By Faith Kwagala
A new report has raised concerns about the growing use of artificial intelligence in healthcare after researchers found that AI systems made errors in many medical diagnoses.The study by Harvard Medical School tested 21 AI systems, including GPT, Gemini and Claude, to see how accurately they could identify diseases from patient symptoms. Researchers found the systems recorded an 80% failure rate when working with incomplete medical information.Experts warned that AI can misunderstand symptoms, miss important details and provide misleading advice, especially during early diagnosis stages.Despite the concerns, researchers say AI still has potential in improving healthcare services and speeding up medical data analysis. In Uganda, institutions like Makerere University and Butabika Hospital are already exploring AI tools for disease detection and mental healthcare support.Health experts, however, insist that AI should support doctors rather than replace trained medical professionals.