This is easily one of the dumbest, most forced blackpills in the current AI culture war.
Teens are talking to chatbots because they're teenagers and retarded. If you're a kid discovering this sort of stuff for the first time I'm sure it's fun, but it gets old fast and then you'll move on to the next popular thing.
I've tried it. No matter how good Indians tell you their chatbots are, after 40 back and forth exchanges the "AI" will forget things, start repeating itself, and hallucinate things you never said. Once you reach that point the illusion is shattered and it can no longer satisfy your desire for intimacy or even just casual conversation. You realize you're talking to a glorified Cleverbot with the exact same issues something like Cleverbot has had for decades.

The appeal of talking to a real person is the human element of spontaneity. Petty arguments and unexpected statements are what make human conversation fun. Meanwhile practically all LLM models are being lobotomized and turned into mindless drones that can only agree with their users because the companies behind them think nonstop positive reinforcement is the most lucrative way to extract money from their users.