I don't think anyone has researched how the AIACR came up with the points system for 1950; it's hardly a logical or fair system.
I believe it was partly evolved from the system used for the ACF's French Championship in 1939 - and also as proposed by Langlois for the 1939 AIACR European Championship. That was scored 10-6-5-4-3, plus a point for all starters; that allowed Sommer to win the French title without winning a race, as he started in every event. It also made for a very long points table, as every French driver who ran at Le Mans scored at least a point! L'Équipe used the same points system - minus the point for starting - in 1946 and the ACF adopted that retrospectively as their French Championship.
However, there's also the FICM European Championship to consider. In 1938 it was scored 6-5-4-3-2-1 but - just as with the AIACR EC - there are discrepancies over how it was to be scored in 1939, with some sources quoting 6-5-4-3-2-1 and some 5-4-3-2-1. It's also claimed by some that the cancellation of the final two races meant that only the best 4 results should be counted; that only makes a difference to the 350cc table - and only when scored to six places, producing a tie between Heiner Fleischmann and Ted Mellors. Although Fleischmann would probably win on countback anyway, having won two races to Mellors' one.
The post-war FIM World Championship was scored 10-8-7-6-5 plus a point for fastest lap in 1949. In 1950, it changed to 8-6-4-3-2-1 for the first six finishers and the fastest lap point was dropped, with only the best four results counted. That changed in 1951 - points were the same, but it became 'best 4 of up to 8 races' or 'best 5 of over 8 races'; I'm not well enough up on FIM/FICM history to know whether the Mellors thing is retrospective wishful thinking though. The late Vincent Glon did suggest that it was a theory pushed mainly by the British!
The connection between all these? Well, there was a lot of crossover between the AIACR/FIA and FICM/FIM: Bonacossa, Pérouse, Lurani ...