A new artificial intelligence tool accurately predicts certain forms of cancer at least three years prior to a diagnosis::undefined
I can accurately predict all forms of cancer 15 years prior to a diagnosis.
Lots of false positives though.
This is not “intelligence.” This is automation.
since when these are mutually exclusive
I think he means algorithms. And since up to a couple years ago, before AI became the new tech buzzword
They use machine learning, which is a subset of artificial intelligence. It’s not just a buzzword here.
Admittedly I’m an amateur, but I consider “automation” to encompass algorithms, heuristics, cron jobs, shell scripts, lambdas, basically anything created to do some steps that we’ve already figured out. As I understand it, machine learning uses statistical algorithms. The article makes the process sound like heuristics, though:
“K-ECAN uses basic information already readily available in the EHR, like patient demographics, weight, previous diagnoses and routine laboratory results, to determine an individual’s risk of developing esophageal adenocarcinoma and gastric cardia adenocarcinoma,” said Rubenstein.
I wouldn’t consider any of that kind of automation to be “intelligence.” Most of the stuff we currently call “AI” is the best form of automation that we can create right now, but it’s still not AGI.
It’s ML using gradient boosting, here’s their paper.
Cool
I agree with you that machine learning does qualify as AI, and that it applies here. That wasn’t the original discussion, though. I was just saying that algorithms did not traditionally qualify as AI, up until the recent trend of calling EVERYTHING AI.
Yeah so much of what is branded as ‘ai’ is misleading imo.
