Topic: Sloan Kettering

Internal documents reveal that IBM's Watson for Oncology often provided "unsafe and incorrect" treatment recommendations. Despite marketing claims that the AI was trained on real patient data, it was actually programmed using hypothetical scenarios created by doctors at Memorial Sloan Kettering. In one instance, the system recommended a drug that could cause fatal hemorrhaging in a bleeding patient, leading to accusations that IBM is experimenting with human health in real-time.