It could be the person was already in a problematic situation with family and friends, and they just need to blame someone or something and don’t want to admit the real problems. Kind of what often happened back in the day with videogames getting blamed for killing humans.

This is not a fair analogy for what is going on here. Video games being blamed harkens back to times when music or other counter cultural media was blamed for behavior. We have a lot of literature which shows that the passive consumption of media doesn’t really affect someone in the ways which they were being blamed.

AI on the other hand, interacts back with you, and amplifies psychosis. Now this is early days and most of what we have is theoretical in nature, based off case-studies, or simply clinical hypothesis [1, 2, 3]. However, there is a clear difference in media itself - the chatbot is able to interact with the user in a dynamic way, and is programmed in a manner by which to reinforce certain thoughts and feelings. The chatbot is also human-seeming enough for a person to anthropomorphize the chatbot and treat it like an individual for the purposes of therapy or an attempt at emotional closeness.

we could end up in a society where everyone undermines real problems in physical world and blames Ai to sideload the question

This is a valid point to bring up, however, I think it is shortsighted when we think in a broader context such as that of public health. We could say the same about addictive behaviors and personalities, for example, and absolve casinos of any blame for designing a system which takes advantage of these individuals and sends them down a spiraling path of gambling addiction. Or, we can recognize that this is a contributing and amplifying factor, by paying close attention to what is happening to individuals in a broad sense, as well as smartly applying theory and hypothesis.

I think it’s completely fair to say that this kid likely had a lot of contributing factors to his depression and ultimate and final decision. There is a clear hypothetical framework with some circumstantial evidence with strong theoretical support to suggest that AI are exacerbating the problem and also should be considered a contributing factor. This suggests that regulation may be helpful, or at the very least increased public awareness of this particular technology having the potential to cause harm to certain individuals.