I, for one, welcomed the compute input. I've seen us suffer from human bias altogether too many times...
I agree that there were years where Bama was hurt because of bias, which computers may have addressed to some degree: '66 and '77. OTOH, computer input would have likely prevented Bama from a claim on the '73 NC and maybe '78. In '73 Bama split with ND after losing to them in the Sugar Bowl. In '78 Bama split with USC after losing to them at Legion Field in decisive fashion. Bama would likely only be able to claim either '64 or '65. So, IMO, Bama would have gained and lost. But there was a definite bias against the South some of it having nothing to do with football, and some because they thought Midwest and West Coast football was superior.
I believe that kind of bias is now. over though there is likely other kinds of more subtle but less egregious bias. In the 60s and even the 70s, the polls were not taken as seriously by the voters as they are now, plus there was little "intersectional football" played, everything was so regional. So there was little reference point for computers. But computers were not a component of rankings at that time and we have no idea if they would have helped.
I just know that now, where human polls or committees differ with computers, the humans come out ahead. I can't remember one time they did not. Not saying they shouldn't be consulted, along with other numerical type data like advanced stats, FPI, etc., but I don't want them having a direct say in the selection. I would assume
that the Selection Committee members do consult them. Really for whatever selection method used, their primary work is greatly dictated by won/loss records, the selectors just evaluate 2-4 teams for the final spot or two. It's not that hard. Only a computer can mess it up badly like 2009 when it would have been UTx, Bama, Cinn, pre P5 TCU. It was Bama/Cinn if UTx doesn't hit a last second 46 yd FG after having a second put back on the clock in their Conf Champ game. That's the kind of disaster they can cause.