OpenAI’s ‘Adult Mode’ to Allow Smut, Not Porn
▼ Summary
– OpenAI’s delayed “adult mode” for ChatGPT will initially support text-based conversations with adult themes, but not the generation of images, voice, or video.
– The feature’s rollout was delayed due to internal concerns, technical safeguards, and a focus on higher-priority tasks, with no new timeline announced.
– Advisers warned the feature could be accessible to children and foster unhealthy emotional dependence, with risks like creating a “sexy suicide coach.”
– OpenAI faces content moderation challenges in allowing NSFW text while blocking harmful scenarios like nonconsensual behavior or child sexual abuse.
– The company’s age-prediction system has misclassified minors as adults, and a text-only approach may help navigate regulations like the UK’s Online Safety Act.
OpenAI’s forthcoming “adult mode” for ChatGPT is designed to permit users to engage in text-based conversations with mature themes, though it will not extend to generating images, voice, or video content. According to a company representative, this feature will facilitate the creation of smut rather than pornography, focusing on written exchanges with adult subject matter. The initiative, first revealed last October, aims to provide verified adult users with access to AI-generated erotica after the company addressed certain safety concerns with its models.
The launch, originally slated for this quarter, has been postponed as OpenAI shifts its attention to other high-priority projects. No new timeline for release has been provided. Internal deliberations and technical hurdles related to implementing effective safeguards have contributed to the delay. An advisory council selected by OpenAI expressed concerns in January, warning that the feature might be accessible to minors and could potentially encourage an unhealthy emotional reliance on the chatbot. One member cautioned that the company risked creating what they described as a “sexy suicide coach.”
Content moderation presents a significant challenge. Sources indicate that OpenAI is working to carefully relax restrictions on not-safe-for-work (NSFW) material while continuing to block more harmful content, such as depictions of nonconsensual acts or child sexual abuse. This balancing act is complex and has slowed development.
Age verification remains a critical issue. An internal system designed to identify and exclude minors was, at one stage, incorrectly classifying them as adults approximately 12 percent of the time. With ChatGPT reportedly attracting about 100 million weekly users under the age of 18, such an error rate could inadvertently expose millions of young people to inappropriate conversations. A spokesperson acknowledged that while their age-prediction technology performs similarly to industry standards, it “will never be completely foolproof.”
Limiting the feature to text may offer a strategic advantage. This approach could help OpenAI navigate regulatory frameworks like the UK’s Online Safety Act, which mandates age verification for pornographic imagery but does not impose the same requirements for written erotic content. This text-focused strategy also distinguishes ChatGPT from competitors offering more visual NSFW experiences. For instance, Elon Musk recently stated that xAI’s Grok is permitted to generate images and videos comparable to what is “allowed in an R-rated movie.”
(Source: The Verge)
