It's already happening, cybersec companies focusing on graymail filters already use bots and AI to clasiffy your email... Making the decision if the message is secure to pass or not.

Personally is due to the toxicity of their dev and socials team, basically if you ask something, you are wrong and they are right, if you recommend other options, you are wrong and they are right. They have been publicly raging war against cybersecurity content creators that dare question some decisions or do honest reviews (OS Is good but has it's but scenarios) ... Once they get better with their PR relationships most of the hate will go away.

I'm sure they have a backup somewhere that they will use to train the AI, but agreed, it is time to leave reddit for good.