I've seem papers before where machine learning was able to compensate for both inaccuracy of measurement and outright falsehoods being reported, and overall was

I've seem papers before where machine learning was able to compensate for both inaccuracy of measurement and outright falsehoods being reported, and overall was far more reliable than humans. Full Disclosure: I just read a paper saying that only 10% of doctors are capable of bayesian reasoning, so I'm in a mood to pick on doctors today. I'll stop now.