Sam Altman Addresses Growing Backlash Over GPT-5’s ‘Sanitised’ Responses

Sam Altman responded to criticism of GPT-5, acknowledging user frustration over its perceived emotional detachment. He admitted retiring older AI models too quickly was a mistake and emphasised safety for vulnerable users. The backlash follows concerns about AI’s role in sensitive advice and its balance between personality and responsibility.

OpenAI CEO Sam Altman has spoken out in response to mounting criticism of GPT-5, the company’s latest AI model, after users voiced disappointment over what they describe as “sanitised” and emotionally distant answers. Many long-time ChatGPT users have compared GPT-5 unfavorably to its predecessors, claiming it lacks depth, nuance, and the human-like warmth they had grown accustomed to.

Some critics have speculated that the changes reflect a cost-cutting approach rather than a leap forward in AI capabilities. Others have expressed frustration at the deprecation of older models, which they argue offered a more engaging and relatable conversational experience.

In a post on X, Altman acknowledged that users have formed unusually strong emotional bonds with past AI models. He admitted that removing older versions too quickly was a mistake, noting that the attachment felt towards AI was “different and stronger” than that associated with past technologies.

Altman also addressed concerns about safety and responsibility, stressing that while most people can distinguish between reality and roleplay, a small but vulnerable percentage cannot. He warned that if a user is mentally fragile and prone to delusion, AI interactions could unintentionally reinforce harmful beliefs.

While affirming OpenAI’s commitment to user freedom, Altman underscored the importance of introducing powerful new technology cautiously. He said billions of people could soon depend on AI for critical life decisions, and that responsibility could not be taken lightly.

Also Read: Bangladesh to Hold General Election in February 2026, Says Election Commission Chief

The discussion comes amid a separate report by the Centre for Countering Digital Hate, which alleged that ChatGPT had provided dangerous advice to teenagers on sensitive topics such as self-harm and substance use. The findings have added fuel to the debate over how AI should balance openness, personality, and safety in its responses.