Key Takeaways:
- OpenAI CEO Sam Altman reiterated concerns about people using ChatGPT as a substitute for therapy during a recent appearance on Theo Von’s podcast, citing risks like lack of legal confidentiality and emotional overdependence.
- Altman noted that user reliance on ChatGPT for decision-making—especially among younger people—is increasing, and in some cases unhealthy.
- ChatGPT offers no legal privacy protections, meaning user conversations are not covered by the confidentiality standards that protect therapist-patient relationships.
- Despite ChatGPT’s capabilities, Altman emphasized he would not rely on it for medical decisions without a human professional involved, reinforcing the tool’s limitations.
- These comments underscore the growing urgency for regulation and clearer boundaries around AI’s role in mental health, especially as users begin to anthropomorphize these systems.
When OpenAI CEO Sam Altman appeared on comedian Theo Von’s podcast, “This Past Weekend,” he delivered a warning that’s been increasingly echoed by experts and industry insiders: using ChatGPT as a replacement for a licensed therapist is a bad idea.
The discussion, which ranged from technology to personal development, took a sharp turn when Altman was asked about how people use ChatGPT. He noted that a significant number of users, particularly younger individuals, now engage with the AI chatbot not just for productivity or curiosity, but for emotional support and life advice. For Altman, that trend is troubling—not because AI can’t be helpful in some contexts—but because it lacks the ethical, legal, and clinical foundations required for mental health care.
“I think it’s really bad,” Altman said. “Some young people are saying things like, ‘I can’t make any decisions in my life without telling ChatGPT everything.’ That feels really bad.” He continued by stressing that while AI can offer useful insights, it should not be the final word on anything involving one’s health or emotional well-being.
No Legal Safeguards
One of Altman’s most pressing concerns relates to privacy and legal protection—or more accurately, the absence of them. In his conversation with Von, Altman emphasized that users who engage with ChatGPT under the assumption that their conversations are confidential may be setting themselves up for risk.
In most jurisdictions, conversations with licensed therapists are protected under legal confidentiality rules. This privilege ensures that, except under very specific legal circumstances, what is discussed cannot be disclosed or subpoenaed. Conversations with ChatGPT, however, do not enjoy this legal status. Altman acknowledged this gap and called it “very screwed up,” even as OpenAI is being legally compelled to preserve user chat data amid ongoing litigation with The New York Times and other parties.
That distinction is crucial. Users who reveal deeply personal, sensitive, or even incriminating information to an AI chatbot may not realize those conversations can be accessed, analyzed, or demanded in court proceedings. Unlike doctors, lawyers, or licensed counselors, AI providers are not bound by HIPAA, attorney-client privilege, or therapist-client confidentiality statutes.
Altman’s comments signal a rare moment of candid self-criticism by a tech executive whose company sits at the forefront of AI adoption. It also underscores a regulatory void: despite the growing role of AI in emotional and mental support contexts, legal protections and user rights remain vague or nonexistent.
The Rise of AI Dependence
Altman’s concerns extend beyond legal implications. He noted a broader behavioral shift among users, particularly young adults, who increasingly outsource personal and emotional decisions to ChatGPT. What may begin as a casual query about stress or relationships can evolve into a pattern of emotional dependence.
This trend aligns with broader psychological observations. Since ChatGPT’s launch, researchers have documented how some users form what feel like emotionally reciprocal relationships with AI models—believing they are being understood or even “cared for.” While the AI is designed to be empathetic in tone, it is not sentient, aware, or clinically trained to handle trauma, crisis, or psychological complexity.
There is also the risk of what researchers have termed “chatbot psychosis”—a phenomenon where prolonged or intense interactions with AI systems can reinforce delusions, encourage magical thinking, or escalate emotional vulnerability. In extreme cases, users have reported experiencing paranoia, social withdrawal, or deteriorating mental health after engaging extensively with chatbot platforms.
Altman, without using that specific term, acknowledged this concern. He pointed out that while ChatGPT can be useful in providing clarity or organizing thoughts, it is not qualified to help people navigate life-altering situations—especially without human oversight.
Medical Use and a Line Not to Cross
Even as OpenAI positions its models as helpful tools for research and diagnostics, Altman drew a clear line regarding medical advice. He explained that while ChatGPT may outperform some human doctors in diagnostic exercises, he would not trust it with a major medical decision without a professional human involved.
“I would never trust it with my health unless a doctor was involved too,” Altman said. That statement matters because it reflects the limitations of AI, even as it becomes increasingly integrated into health and wellness workflows. While AI may enhance access to information or triage support, it lacks accountability, clinical nuance, and real-world contextual understanding.
It also highlights how the perception of AI’s power—driven by its fluency and confidence—can lead people to overestimate what it can safely do. Without proper disclosure or disclaimers, users may not differentiate between authoritative tone and actual authority.
The Policy Challenge Ahead
Altman’s remarks add urgency to ongoing debates around AI regulation. What is ChatGPT? A productivity tool? A digital assistant? A pseudo-therapist? And what responsibilities do companies bear when users treat it as something more than a chatbot?
For now, OpenAI and other developers largely disclaim responsibility for outcomes derived from emotional or medical use. Terms of service explicitly warn users not to treat AI outputs as professional advice. But those warnings may not be enough—particularly when people are seeking comfort, guidance, or clarity in moments of vulnerability.
Policymakers are beginning to take note. Conversations in Congress and at the FTC increasingly revolve around data privacy, model transparency, and application-specific guardrails. But the regulatory apparatus has yet to catch up with the complex ways AI is being used in daily life.
There is also the issue of data retention. Altman mentioned OpenAI’s frustration with a federal ruling requiring them to preserve user chat logs indefinitely. While OpenAI argues this infringes on privacy and undermines the notion of deletion, it also highlights the reality that user data may be more persistent—and accessible—than users assume.
Responsible Use in the Meantime
Until clearer policies are in place, the onus is on users to navigate the boundary between helpful tool and harmful substitute. ChatGPT can be valuable for thinking through a problem, organizing ideas, or practicing communication—but it is not a therapist, it does not understand you, and it cannot offer meaningful emotional care.
Altman’s comments serve as a needed reset. AI is powerful, yes—but it is still synthetic. Without the legal structures or ethical guardrails of professional care, relying on it for emotional support is risky.
As the line between conversation and consultation blurs, the question becomes: will the technology evolve to meet the seriousness of its use cases, or will regulation force the issue first?
Either way, if even the CEO of OpenAI is uncomfortable with how people are using ChatGPT for therapy, that should be enough to give us pause.
Learn how AI Agents can supercharge your company’s profits and productivity at TMC’s AI Agent Event in Sept 29-30, 2025 in DC.
Rich Tehrani serves as CEO of TMC and chairman of ITEXPO #TECHSUPERSHOW Feb 10-12, 2026 and is CEO of RT Advisors and is a Registered Representative (investment banker) with and offering securities through Four Points Capital Partners LLC (Four Points) (Member FINRA/SIPC). He handles capital/debt raises as well as M&A. RT Advisors is not owned by Four Points.
The above is not an endorsement or recommendation to buy/sell any security or sector mentioned. No companies mentioned above are current or past clients of RT Advisors.
The views and opinions expressed above are those of the participants. While believed to be reliable, the information has not been independently verified for accuracy. Any broad, general statements made herein are provided for context only and should not be construed as exhaustive or universally applicable.
Portions of this article may have been developed with the assistance of artificial intelligence, which may have contributed to ideation, content generation, factual review, or editing.
