Sam Altman sounds alarm over AI dependence among teens, calls it 'bad and dangerous'
According to Sam Altman, this mindset is not a fringe phenomenon but a “widespread pattern” among young users, particularly those in their teens and twenties.
ADVERTISEMENT
As generative AI tools like ChatGPT become increasingly embedded in our everyday lives, OpenAI CEO Sam Altman is issuing a sobering warning: don’t let AI make your life decisions.
Speaking at a Federal Reserve conference on banking and technology, Altman revealed that a troubling number of teenagers and young adults are turning to ChatGPT not just for homework help or career tips—but for deeply personal guidance.
“ChatGPT knows me, it knows my friends — I’ll just do what it says,” Altman quoted users as saying. “That feels bad and dangerous.”
According to Altman, this mindset is not a fringe phenomenon but a “widespread pattern” among young users, particularly those in their teens and twenties. For them, ChatGPT is not just a chatbot — it’s become a life coach, confidant, and in some cases, a kind of operating system for their daily existence.
AI As Life Advisor Altman described distinct patterns in usage across age groups. While older users treat ChatGPT like an advanced search engine, college students often rely on it to structure their entire digital and academic lives. From writing emails to organizing thoughts and even journaling — for some, it’s a daily companion that feels deeply personal.
This raises ethical red flags. The emotional connection users build with AI — especially when it starts “knowing them better” than people around them — could have long-term implications on autonomy, mental health, and critical thinking.
Survey: Teens Trust AI More Than You’d Think Altman’s comments align with recent findings by Common Sense Media, which surveyed 1,060 teens aged 13–17. The report found that 72% had used an AI companion at least once, 52% used AI tools regularly each month, nearly half said they trusted the advice they received, younger teens (13–14) showed the highest levels of trust.
These numbers underscore just how comfortable Gen Z is becoming with AI — and how easily that comfort could evolve into dependence.
Experts Agree: Overtrust Is a Risk Altman’s warning echoes those of other AI pioneers, like Geoffrey Hinton, the so-called “Godfather of AI,” who recently confessed that even he finds himself trusting AI outputs more than he should — despite knowing their limits.
“Even the best models can make mistakes on basic logic problems,” Hinton noted in an interview, adding that trust in AI must be earned through transparency and accountability — not blind faith.
Deepfakes & Cybercrime: The Broader Threat Altman also spoke about AI misuse beyond personal overreliance, warning banks and financial institutions of an impending “fraud crisis.” He called out the continued use of voice-based authentication — which can now be hacked with near-perfect voice cloning — as “crazy.”
He further cautioned that realistic video deepfakes could soon compromise facial recognition systems as well.
“We’re approaching a fraud crisis,” Altman said. “It’s time to seriously rethink how we protect identity in the AI era.”
