

From what I understood, it’s an unintended consequence of accessibility rules coming into place.
It’s a fully intended consequence of DRM refusing to adapt to said accessibility rules. Closed ecosystems make DRM easier, which was always the goal for publishers.
You may want to look up the study “Speaker sex and perceived apportionment of talk” for a potential explanation of why this could be happening.
Basically, psychologists did a study where they asked participants to rate excerpts from a play. They started by attempting to control for male and female “role” bias from the script itself; They had university students read the scripts (with “A” and “B” listed as the speakers’ names, gendered pronouns swapped for neutral pronouns, etc) and try to intuit the sex of the characters in the play. So this gave them a baseline on the socially perceived gender of the roles in the script. So if one role was filling a more traditionally feminine or masculine role, had more fem/masc speech patterns, etc, this part of the study was designed to check for that.
Next, they had actors perform the script, and took some recorded excerpts to play for participants. The excerpts had a male and female actor, and the participants needed to rate how long they believed the excerpt was, and how much they believed each actor spoke, from 0-100% of the conversation. So for instance, if they believed the female actor spoke 40% of the time, they would list 40 for her and 60 for the male actor.
Virtually every single participant (both male and female) over-estimated the female actor’s participation to some degree. Female participants were closer to reality, but male participants were pretty far off. Some of the male participants began saying the woman was an equal contributor when she was only speaking 25-30% of the time. Interestingly, these numbers were closer to reality (not totally accurate, but closer) when they flipped the script (literally) and had the actors play the opposite roles. So the female actor was now playing the “male” (determined by the earlier script reads) part of the script. So societal role expectation does play some part in the determination… But it’s not the entire reason.
It could be a large part of why so many terminally online men pipe up about “feminism is ruining my hobbies” whenever more than a token woman is added to media. Because many men genuinely feel like women are an equal contributor when they’re only a small fraction. Does it excuse the behavior? Absolutely not. But it could at least begin to explain it.