Thread by @kegill

Do we need more examples of how “AI” is not ready for prime time? You’d think pattern recognition would be the easiest ...

https://twitter.com/kegill/status/1022548958273576961

Here’s MD Anderson and IBM’s Watson from 2017: failure to “expedite clinical decision-making around the globe and match patients to clinical trials.”

healthnewsreview.org/2017/02/md-and…

More on how @IBMWatson has failed cancr patients from @Gizmodo yesterday, citing paywall @statnews

gizmodo.com/ibm-watson-rep…

The failure rate - often inaccurate recommendations - is unacceptable given that “Watson for Oncology is trained to help oncologists treat 13 #cancers and is being used by 230 hospitals around the world, and has ‘supported care for more than 84,000 patients’.”

“This product is a piece of shit,” a doctor at Florida’s Jupiter Hospital said to @IBM, according to the documents reviewed by Stat.

Perhaps it was the teacher’s fault-surgeons at Memorial Sloan Kettering gave @IBMWatson hypothetical patient data (“synthetic” case data) rather than real patient data.

And @MDAndersonNews reported problems with Watson for Oncology last year.

Babylon, an AI app, “claims to be able to diagnose medical conditions better than human doctors.” But a doctor claims it does not “properly identify heart attacks.”

July 2018 report from the UK

telegraph.co.uk/technology/201…