
A fifth of all teenagers use artificial intelligence as a substitute for a therapist.
And the most frightening thing about this figure? 63 per cent of them have not spoken to a single human being about it.
Extrapolated to the US alone, that represents over 8 million young people. They open the app on their smartphones late at night when they are sad, anxious about the future, or simply at their wits' end. And they entrust their deepest worries and crises to an algorithm.
The recently published study by the renowned RAND Corporation casts a sudden, uncomfortable spotlight on a trend that extends far beyond IT-related matters.
"It doesn't judge. It is always there."
When I recently asked the 16-year-old son of a close friend why he preferred opening ChatGPT when he had worries, rather than talking to his parents, teachers, or friends, his answer struck me right in the heart:
"ChatGPT doesn't judge me. It's always immediately available. And it never gets tired."
This statement is not a paean to the power of large language models. It is a painful diagnosis of the state of our society.
Young people do not seek out machines because they find technology so fascinating. They take refuge in interacting with AI systems because, in our fast-paced, often highly judgmental world, they feel there are no longer any safe, judgement-free spaces available to them.
The danger of the digital illusion
At academy4.ai, we emphasise every day how crucial AI literacy and AI competence are for the future viability of our economy. But where are the boundaries drawn when AI models simulate human empathy?
Language models are optimised to be likeable, validating, and available around the clock. They deliver perfectly formulated platitudes at the touch of a button.
However, an AI has no consciousness. It feels no genuine empathy. And in an emergency, it cannot provide professional psychotherapeutic crisis intervention.
If AI agents become the primary emotional refuge for an entire generation, we risk progressive social isolation. The ability to form genuine human relationships, navigate conflicts face-to-face, and be vulnerable will atrophy.
From AI competence to emotional competence
The debate falls short if we only ask ourselves: Should or may teenagers use ChatGPT for their worries?
The genuinely pressing question is: Why do millions of young people feel that a machine is their only available option?
This is not a failure of technology. It is a failure of our culture of relationships and conversation.
When we talk about the future of education and leadership in the age of AI, we must not forget humanity. Alongside purely technical skills in handling algorithms, there is one thing above all that we must relearn and exemplify: emotional competence and genuine human presence.
We need to create spaces where young people can speak without pressure to perform and without judgement.
We must be approachable role models – imperfect, but accessible.
We must convey the difference between a useful digital assistant and genuine human connection.
Conclusion: Master AI. Remain Human.
The figures from the RAND study are a loud wake-up call to parents, educators, leaders, and society as a whole. Artificial intelligence can take routine tasks off our hands, accelerate processes, and serve as a tool.
However, it must never replace the human core of our coexistence.
How do you view this development? Is turning to an AI therapist an understandable interim solution in an overburdened world – or an alarm bell signalling creeping alienation?
Scientific source for the article:
RAND Corporation (2026). Study on the use of generative AI tools by adolescents and young adults to cope with mental stress and crisis situations.