Is machine learning overhyped?

What's the area of chemistry in which you think machine learning has the potential to have the biggest impact? "In my experience, the availability of information is so overwhelming that it is hard to make sure that the conclusions that we obtain from the literature search are not biased. I think that machine learning will help to obtain statistically unbiased bibliographic information."--Samuel "It has the potential to see trends in data that humans tend to overlook. The tools provided by machine learning will aid the chemist in decision-making by providing suggestions that may spark new ideas."