Chat-GPT becomes Sex-GPT for verified adults (twitter.com)

OpenAI has announced plans to permit erotic content on ChatGPT, marking a notable shift in the platform’s content policy. Moving from an across-the-board ban to a controlled allowance for adult sexual content signals a balancing act between user demand for expressive capabilities and the company’s safety and legal obligations. For users this likely means opt-in settings, stricter age verification and clearer consent and consent-checking requirements; for platforms and regulators it raises jurisdictional and moderation questions about illegal or exploitative material. Technically, enabling erotica forces changes across the model stack: updated content policies, retrained or fine-tuned models with revised reward signals (RLHF) to produce allowed material while refusing disallowed types, improved NSFW classifiers, and more granular safety layers to detect edge cases (e.g., minors, non-consensual scenarios, trafficking). It also impacts dataset curation, evaluation metrics, adversarial-prompt defenses, and logging/traceability for compliance. For the AI/ML community this is significant because it sharpens research priorities around robust content filters, contextual understanding of consent and age, mitigations against misuse, and transparent policy-to-model translation — all necessary to safely expand generative model capabilities into legally and ethically fraught content domains.