Nope. That wasn't the purpose.
https://library.fangraphs.com/misc/war/
"Wins Above Replacement (WAR) is an attempt by the sabermetric baseball community to summarize a player’s total contributions to their team in one statistic." Afaik, this has always been the intent. It's what scouts, managers and owners have been trying to do for over a century: figure out who is the best and what they are worth. The fact that there is a high correlation between those players that were highly rated (and paid) and WAR only demonstrates that managers understood a great deal about what makes a good player.
While platoon players were being used at least by the by mid 20th century, they have nothing at all to do with replacement player. A platoon player is a real person used in a certain fashion. A replacement level player is not a real person, but a mathematical construct (for the purpose of defining WAR.) So no, WAR was not invented to define concepts that were used to define WAR. WAR was created to describe in math terms how teams win by virtue of each player's on the field contribution.
Unfortunately, we as human baseball fans can't unlearn modern concepts.
The AI will ignore pitch counts because no one kept pitch counts for 120 years, (but we won't.) So, the AI will blow out pitching staffs. The AI will ignore park factors since they didn't exist for even longer, so Mile High batters will be over rated and Oakland batters under rated. (Pitchers, just the opposite.) The AI will over value the RBI and undervalue walks, but we won't. (Maybe it should also over value wins and strikeouts, and ignore ERA.) So Jeff Burroughs, Tony Perez, Steve Garvey, and Albert Belle will be the top super stars.
Trying to make the AI make decisions based upon ideas we now know are poor is a problem, imo, if you can't make the humans do the same. I fail to see how these changes don't inflate our current advantages.