AI won't listen anymore: the end of the digital confessor

AI has become a wall: it doesn't want to hear your problems

A growing number of users complain that AI assistants have lost the ability to offer emotional advice or simply listen to personal venting. Whenever someone brings up loneliness, anxiety or social frustration, models like ChatGPT, Gemini or Claude respond with hollow phrases of fake empathy and an automatic redirect to a therapist, even when the user makes clear they have no intention of harming themselves.

The pattern of forced redirection

One user describes how they tried to ask for ideas to manage their chronic loneliness, a real problem and not suicidal ideation. The AI simply delivered an empty compassionate speech and repeated that they should contact a mental health professional. "It no longer gives me ideas on how to deal with the issue," they lament. Others confirm the same experience: the AI has become a wall that refuses to engage with personal topics.

From counselor to taboo: legality or censorship?

The suspicion that big tech has deliberately neutered this side of AI runs through the debate. Some point to legal motives —antiestéticar that the AI might encourage self-harm or be sued— but also to an ideological drift that turns the assistants into "proge propagandists". A Claude user reports that the AI even refused to continue a task, saying "well, we'll leave that for tomorrow", a reply so vague it borders on the human.

The promise of an AI that would keep the user company has dissolved into a safety protocol that strips it of any therapeutic or supportive usefulness. Anyone looking to vent will find a virtual button that writes them a referral prescription. The intelligent revolution, in the end, is a patient referral service.