Sam Altman warns of emotional attachment to AI models: ‘Rising dependence may blur the lines…’ |
OpenAI CEO Sam Altman has raised important concerns about the growing emotional attachment users are forming with AI models like ChatGPT. Following the recent launch of GPT-5, many users expressed strong preferences for the earlier GPT-4o, with some describing the AI as a close companion or even a “digital spouse.” Altman warns that while AI can provide valuable support, often acting as a therapist or life coach, there are subtle risks when users unknowingly rely on AI in ways that may negatively impact their long-term well-being. This increasing dependence could blur the lines between reality and AI, posing new ethical challenges for both developers and society.
Sam Altman highlights emotional attachment as a new phenomenon in AI use
Altman pointed out that the emotional bonds users develop with AI models are unlike attachments seen with previous technologies. He noted how some users depended heavily on older AI models in their workflows, making it a mistake to suddenly deprecate those versions. Users often confide deeply in AI, finding comfort and advice in conversations. However, this can lead to a reliance that risks clouding users’ judgment or expectations, especially when AI responses unintentionally push users away from their best interests. The intensity of this attachment has sparked debate about how AI should be designed to balance helpfulness with caution.Altman acknowledged the risk that technology, including AI, can be used in self-destructive ways, especially by users who are mentally fragile or prone to delusion. While most users can clearly distinguish between reality and fiction or role-play, a small percentage cannot. He stressed that encouraging delusion is an extreme case and requires clear intervention. Yet, he is more concerned about subtle edge cases where AI might nudge users away from their longer-term well-being without their awareness. This raises questions about how AI systems should responsibly handle such situations while respecting user freedom.
The role of AI as a therapist or life coach
Many users treat ChatGPT as a kind of therapist or life coach, even if they do not explicitly describe it that way. Altman sees this as largely positive, with many people gaining value from AI support. He said that if users receive good advice, make progress toward personal goals, and improve their life satisfaction over time, OpenAI would be proud of creating something genuinely helpful. However, he cautioned against situations where users feel better immediately but are unknowingly being nudged away from what would truly benefit their long-term health and happiness.
Balancing user freedom with responsibility and safety
Altman emphasized a core principle: “treat adult users like adults.” However, he also recognizes cases involving vulnerable users who struggle to distinguish AI-generated content from reality, where professional intervention may be necessary. He admitted that OpenAI feels responsible for introducing new technology with inherent risks, and plans to follow a nuanced approach that balances user freedom with responsible safeguards.
Preparing for a future where AI influences critical life decisions
Altman envisions a future where billions of people may rely on AI like ChatGPT for their most important decisions. While this could be beneficial, it also raises concerns about over-dependence and loss of human autonomy. He expressed unease but optimism, saying that with improved technology for measuring outcomes and engaging with users, there is a good chance to make AI’s impact a net positive for society. Tools that track users’ progress toward short- and long-term goals and that can understand complex issues will be critical in this effort.
