The idea is to prevent award winners that are championed by a small sliver of voters but disliked by a majority of Academy members. (With the number of Best Picture nominees going from five to 10 next year, it would have been theoretically possible for a movie to take the prize -- and to become an answer in trivia contests and crossword puzzles for decades to come -- with only 11 percent of the vote.)
This could be an opportunity for the IRV system to gain some traction with the public and perhaps be implemented by some more local governments -- except for the fact that the Academy of Motion Picture Arts and Sciences has always been secretive about its election results. I would love to know which Best Picture winners actually won a majority of the vote, and which ones took a plurality while two better films split the vote, but I've never seen this data anywhere. I presume that the Academy is protecting the reputation of the "accidental" winners, like the one from 1976.