What happened

IBM Watson for Oncology, developed in collaboration with Memorial Sloan Kettering Cancer Center and launched in the early 2010s, promised to revolutionise cancer care by providing treatment recommendations based on patient data and medical literature. 

However, reports revealed that the system frequently suggested unsafe or incorrect treatments, relied on limited datasets, and struggled to handle real-world clinical complexity. In one instance, it suggested a drug that could cause severe bleeding for a patient already experiencing heavy hemorrhaging. 

Hospitals in the U.S. and India that implemented the system reported minimal practical benefit, and in some cases, physicians had to override or ignore its recommendations, highlighting a gap between marketing claims and actual performance. 

The failures potentially endangered patients, misallocated hospital resources, and undermined trust in AI in medicine. 

A 2017 audit by the University of Texas System found that the project had spent USD 62 million but had failed to deliver a functional product. The audit also raised concerns about the project's governance, management, and the lack of clear goals and objectives.

The controversy surrounding the project led to a number of high-profile departures, including the resignation of the head of MD Anderson's IT department and the departure of several top executives from IBM Watson Health.

The project was ultimately shut down in 2018, with MD Anderson writing off the entire investment.

The failure of the project was seen as a major setback for IBM Watson Health, which had been touted as a leader in the field of AI-powered healthcare.