That's exactly what they've been doing! Or, perhaps more correctly, they've let their "algorithms" run amok without doing any quality control on the final results, thereby turning modest warming into "scary" warming. Whether this was intentional or not is debatable (see https://en.wikipedia.org/wiki/Hanlon's_razor), but it certainly looked that way to me at the time.

About ten years ago, after some remarkably incorrect statements about local heat levels made by the folks at NOAA or wherever, and then repeated (but very quickly retracted) by our own local weather folk, I went to the trouble of doing a bit of a deep dive on the HadCrut3 data sets. And what I found there was appalling. Other folks did more comprehensive deep dives and found far more problems with it than I had.

I could only do what I did at the time because some authority (I forget who) had finally forced the entire data set to be released to the public, over the strenuous objections of the folks who administer that data. Much of that data has since been hidden again, though, from what I can tell.

That was the older HadCRUT3 data set. The newest one is HadCRUT4, but I haven't done any kind of deep dive on that. (Without digging further into the matter I don't even know if there is enough data publicly available to do the kind of analysis on it that I did on the earlier set.) But other people have, and have noted serious problems with it, and have even written entire books about that. The short version, though, is that there is an awful lot of "data" out there which they're using that is mostly just made up, and what is genuine has been adjusted so much that it can no longer really be considered "data" at all.