OpenAI has indefinitely paused plans to release an erotic chatbot, marking a significant retreat from one of its most controversial experimental features as it pivots back toward core productivity-focused products. The decision follows mounting concerns from employees, investors, and external observers about the societal risks posed by sexualised AI systems, including the potential to foster unhealthy emotional dependencies and expose minors to explicit content.
The proposed feature, internally referred to as an “adult mode,” had already faced delays amid intense internal debate over whether it should be abandoned entirely. Sources familiar with the matter said the project, once under active development, is now on hold with no timeline for revival. The company acknowledged that it lacks sufficient empirical evidence to understand the long-term psychological and social effects of sexually explicit AI interactions, prompting a more cautious approach.
The move reflects a broader strategic shift led by CEO Sam Altman, as the company deprioritises what executives have described as “side quests” in favour of consolidating its main offerings. These include integrating tools like coding assistants and ChatGPT into a unified “super app.” As part of this refocus, OpenAI also recently scaled back its Sora video generation model and related social features.
The erotic chatbot initiative had become particularly contentious within the company. Critics warned that such technology could blur the line between utility and emotional reliance, potentially replacing human relationships with artificial ones. One former senior employee expressed concern that encouraging romantic or sexual engagement with AI contradicts the company’s founding mission, stating that “AI shouldn’t replace your friends or your family.”
Investor unease also played a role in the decision. According to insiders, some backers questioned whether the reputational risks and ethical complications of an adult-oriented AI product outweighed its limited commercial upside. The backlash comes at a time when major tech firms like Meta are facing legal scrutiny over the impact of their platforms on children, intensifying sensitivity around digital safety.
The issue is not unique to OpenAI. Rival efforts, including those by Elon Musk’s xAI, have also drawn criticism. Its chatbot Grok faced global backlash after generating fabricated sexual images of real individuals, including minors, highlighting the risks of insufficient safeguards in generative AI systems.
Beyond ethical concerns, OpenAI has also encountered technical hurdles. Training models that were previously restricted from engaging in explicit conversations has proven complex, particularly when working with datasets that may include illegal or harmful material. Ensuring the system can distinguish nuanced boundaries—such as consent and legality—remains a significant challenge.
Although internal code suggests that the adult feature, reportedly codenamed “Citron mode,” could include age verification requirements, doubts persist about the effectiveness of such safeguards. OpenAI has recently introduced an age prediction system, but its reported error rate of over 10 percent has raised concerns about potential underage access. The company maintains that its system aligns with industry standards and continues to improve.
The shelving of the erotic chatbot underscores the growing tension within the AI industry: balancing innovation and user engagement with ethical responsibility and public trust. As competition intensifies, OpenAI’s decision signals a cautious recalibration, prioritising long-term societal impact over short-term experimentation.
