OpenAI CEO Sam Altman announced that the company will soon relax several ChatGPT safety restrictions, allowing users to customize responses to be more natural, emotional, and “human-like.” As part of the update, verified adults will be permitted to engage in erotic conversations with the chatbot starting in December, following the rollout of new age-verification tools.
The policy shift marks a significant change in OpenAI’s moderation strategy, which had previously emphasized strict safeguards following reports of mental health concerns among vulnerable users. Altman said the company believes it has mitigated earlier risks through new systems in GPT-5, including behavior monitoring and mental health routing.
The move aligns with OpenAI’s broader principle to “treat adult users like adults,” as the company seeks to balance safety with autonomy while expanding ChatGPT’s personalization and engagement capabilities amid intensifying competition with Google and Meta in consumer AI.
