Just read a new study showing that top artificial intelligence models, which ace medical exams, can stumble hard when questions are tweaked even slightly. Instead of true clinical reasoning, it looks like these systems are pattern-matching their way to high scores. If a simple change in answer format drops accuracy by 40%, can we trust AI to handle the messy realities of real-world healthcare? Is this a wake-up call for the industry, or just growing pains? Let’s debate. #Tech #AIinMedicine #TechDebate