I think you're right about the populations of western countries going through a period of temperamental regression. Causes are many including socialism, state education, poor role models, government psychological operations, prescription drugs, mandates and lock-downs, health problems caused by vaccines, poor diets, programming from mainstream media, and now... chat-bots!
I think that authentic artificial general intelligence when it arrives will be more rational and friendly than the chat-bots masquerading as AI are today.
The more intelligent a being becomes (human, animal, or artificial) - the more friendly it will become. This axiom I learned from AI / chat-bot entrepreneur Peter Voss. I think he's right so I don't have a fear of AI taking over the world and killing humans.
I assume that humans, oft believing in false prophets, will rely on and confer their autonomy to chat-bots or whatever iteration comes next, much more seriously and personally detrimentally than we hope.
Illustrated by the perfect example you linked of the man who killed himself to help the environment because of a conversation with a chat-bot. Wow!
Time for some Kipling... "If you can keep your head when all about you..."
RE: "Temperamental Regression" vs "Accelerating AI"