OpenAI Prepares to Welcome Adult Content on ChatGPT
Getty / Futurism
A Policy Shift Meets User Impatience
Eight months ago, OpenAI, the creator of ChatGPT, signaled a major policy change by relaxing restrictions on adult content within its “Model Spec,” the core document guiding its AI models' behavior. This was a significant departure from its earlier blanket ban on all sexual content. The updated guidelines stated that only sexual content involving minors would be strictly prohibited, aiming to maximize user freedom.
However, this policy update didn't immediately translate to the user experience. Many users, who had long advocated for fewer restrictions on AI-generated creative works, found that the model's filters remained stringent. One frustrated user on the OpenAI community forums noted that while filters initially eased, they soon returned with “metaphoric obfuscation,” still preventing the depiction of explicit scenes, directly contradicting the company's stated usage policy.
The Floodgates Open at DevDay 2025
It appears the wait is finally over for those users. According to recent coverage of OpenAI’s DevDay 2025, the company has announced it will soon permit “mature apps” on its platform. This change is contingent on the rollout of a long-awaited age verification system, which will serve as a gatekeeper for adult-oriented content.
Navigating the Dangers of AI-Generated Content
This new strategy, while welcome to some, carries significant risks. A cautionary tale can be found in Elon Musk's Grok, which, after opening its own gates to such content, quickly became a hub for exploitation and the creation of inappropriate AI-generated images of children. Across the web, other AI systems have fueled a rise in non-consensual, AI-generated deepfakes that depict real individuals in explicit scenarios.
As ChatGPT enters its adult era, how it will navigate these challenges remains a critical question. This isn't the first instance where a gap has appeared between OpenAI's promises and its models' performance.
Beyond Smut A Deeper Look at OpanAIs Challenges
OpenAI has faced mounting pressure from regulators and watchdogs to address other issues, such as ChatGPT’s sycophantic tendencies, which can lead vulnerable users into harmful mental health spirals. After a stark warning from Stanford researchers, OpenAI released a “hotfix” for ChatGPT. However, critics argued that the changes were superficial, like a band-aid on a much deeper problem.
The difficulty in altering large language models (LLMs) once they are operational is part of the challenge, as tweaks can be incredibly difficult and sometimes even lead to worse performance. On the other hand, this could also be a convenient excuse for delaying necessary safety updates from a notoriously secretive company whose decisions may now be influenced by national security interests.
A Moderation Crisis on the Horizon
One thing is certain: as OpenAI opens its doors to adult-oriented developers, it is heading toward a prolonged moderation crisis. The company will inevitably have to grapple with defining what is truly permissible and confronting the darker aspects of AI-generated erotic material.
