Kevin B. Screwed

Coming late (an entire day!) to the hullabaloo surrounding YouTube's blinkered decision to disable Kevin B. Lee's account due to alleged copyright infringement. Having served as guest commentator on two of Kevin's outstanding critical video-essays (Un coeur en hiver and El Cid), I'm not a wholly disinterested party in this matter—but still, come on. How difficult would it be to tweak the bot that crawls around looking for digital watermarks or signatures or whatever so that it recognizes which videos likely constitute fair use? Kevin's essays typically employ multiple short clips from the film in question (and sometimes other sources), with no particular snippet running more than perhaps 60-90 seconds, usually much less. I'm not a programmer, but surely there must be an easy and automated way to distinguish that sort of thing—and other clearly creative/analytical/non-regurgitative instances, e.g. those dorky fanboy remixes of favorite TV shows (the appeal of which I personally cannot fathom, but god bless 'em imo)—from the folks who are just posting Leatherheads 1/10, 2/10, 3/10, etc. Couldn't the crawler determine how much of the copyrighted material is contiguous? If we can invent that insane iPhone app that tells you what song it just heard a 20-second snippet of, we can do this.