Watson, IBM’s highly touted artificial intelligence (AI) platform, might not be ready to make its rounds, according to an investigative report from STAT. The news outlet reviewed internal IBM documents that showed Watson often recommended unsafe treatment advice and incorrect recommendations, all while IBM was promoting the AI product to healthcare providers.
The incidents mentioned in the IBM documents happened in 2016, with incomplete training cited as the reason for the shortcomings. According to the STAT report, IBM researchers and physicians at the Memorial Sloan Kettering Cancer Center trained the AI software with hypothetical cases, instead of situations with real patient data.
IBM responded to the STAT report by claiming it presented an out-of-date snapshot of Watson.
“Today, Watson for Oncology is trained to help oncologists treat 13 cancers,” IBM said in a statement. “Our oncology and genomics offerings are used by 230 hospitals around the world and have supported care for more than 84,000 patients, which is almost double the number of patients as of the end of 2017. … We have learned and improved Watson Health based on continuous feedback from clients, new scientific evidence and new cancers and treatment alternatives. This includes 11 software releases for even better functionality during the past year, including national guidelines for cancers ranging from colon to liver cancer.”
This isn’t the first time questions have been raised about the performance of IBM’s AI platform. Reports arose in September 2017 that Watson struggled with interoperability and data collection—problems that can be found in most health IT products.
But critics argued IBM was aiming too high—setting sights on curing cancer rather than improving treatment and diagnosis.