- Updated: March 20, 2026
- 7 min read
ChatGPT Adult Mode and Intimate Surveillance: Risks and Implications
Skip to main contentSecurityPoliticsThe Big StoryBusinessScienceCultureReviewsChevronMoreExpandNewslettersPodcastsVideoLivestreamsMerchSearchSearchCommentLoaderSave StorySave this storyCommentLoaderSave StorySave this storyJulie Carpenter doesn’t blame human users at all for anthropomorphizing chatbots.The longtime human-AI interaction expert and author of The Naked Android sees generative AI tools, like ChatGPT, as intentionally designed to elicit these reactions from users through highly personalized memories and interactions that replicate social experiences.These one-sided connections may become even more complicated, and potentially disastrous, when AI-generated erotica is added as an option. Carpenter is apprehensive about OpenAI’s upcoming plans to allow adults to sext with ChatGPT.“I’m for people exploring their sexuality, but I want it in a way that’s safe for them as well as fun,” she says. “People have to be very aware that there’s a surveillance aspect to the data.”OpenAI first hinted at allowing adult users to generate erotica using ChatGPT two years ago in an official document describing how its models are designed to behave. While the company still seems to be planning its eventual release, the timeline is murky.In recent months, OpenAI’s advisory council of outside experts expressed concerns about the risks of this imminent release, including the possibility of a “sexy suicide coach,” according to a report from The Wall Street Journal. OpenAI declined to comment when reached by WIRED.People talking dirty to machines, like ChatGPT, is nothing new.“They want to monetize something they see that people are going to try and do anyway,” says Kate Devlin, a professor of AI and society at King’s College London, whose research involves digital sex.What’s different here is OpenAI’s potential embrace of smut on a mainstream platform. Erotic chatbots are a revenue-driving business, but they are often relegated to niche sites with unremarkable names or Elon Musk’s Grok, known for nonconsensual deepfake images and going into a “MechaHitler” mode.When the adult mode is released, it’ll be interesting to see what exactly ChatGPT will remember about the sexual proclivities of users. Over the past couple of years, OpenAI has continued to improve the chatbot’s memory feature, which logs user preferences and responds with personalized outputs. Are you a vegan? ChatGPT won’t recommend a steakhouse when asked for restaurant ideas. Want to go on a hike?It knows you’re in San Francisco from past chats, and will suggest reasonably close trails, even if that location context is not included in the prompt.But, what happens when this type of data logging and hyper-specific personalization meets the horny sexting habits of millions of adult ChatGPT users? Erotic interactions with ChatGPT will further heighten the stakes for user privacy.Rather than remembering your dietary preferences or what type of movies you enjoy the most, ChatGPT could recall how much you fantasized about furry porn, threesomes at the beach, and piss play, as well as tailor its daily responses with these sexual details in mind. It remains unclear how OpenAI plans to handle these mature conversations.People who choose to interact with the smut version of ChatGPT at launch may find solace in the “temporary chats” feature.When enabled, these conversations are not shown in a user’s log history or used to improve OpenAI’s models.But these chats are not as ephemeral as they may feel for users. “For safety purposes, we may still keep a copy for up to 30 days,” reads OpenAI’s website. There’s also a disclaimer included on the site page claiming that “data retention for certain services may be affected by recent legal developments.”“You’re sharing your most intimate sexual thoughts because you’re lost in the moment.”Julie Carpenter, author of The Naked AndroidSmutty AI chats enable a kind of intimate surveillance that goes beyond someone’s porn-watching history or who they’re subscribed to on OnlyFans. The details are more specific. The interactions can escalate over months to more and more extreme fantasies. The kind of fantasies you never initially considered or would ever express to a human partner.While the interactions may feel private, the data is still a major risk.If an account is hacked, a security breach occurs, or a government official demands access to your chat logs, troves of highly sensitive and sexual conversations could be exposed in ways the user never imagined.Information about users’ ChatGPT interactions has previously leaked or been unintentionally made public. OpenAI briefly took ChatGPT down in 2023 when a bug exposed the titles of some users’ chat history.“It’s also possible the first message of a newly created conversation was visible in someone else’s chat history if both users were active around the same time,” read OpenAI’s statement at the time. Last year, multiple ChatGPT interactions were unintentionally made public by users who didn’t realize their sharing settings would index the chats on Google Search for anyone to discover.“You’re sharing your most intimate sexual thoughts because you’re lost in the moment,” Carpenter says.“You’re vulnerable in that way because you’re under the impression that you are in this cool, transformative, almost co-constructive creative space that seems like a great way to practice your sexual fantasies.”CommentsBack to topTriangleYou Might Also LikeIn your inbox: The week’s biggest tech news in perspectiveThis popular pro-Trump X account is apparently run by a White House stafferBig Story: The five big ‘known unknowns’ of Trump’s war with IranThe system that intercepted Iran’s missiles over the UAEListen: The Pentagon vs. “woke” AnthropicReece Rogers is WIRED’s service writer, focused on explaining crucial topics and helping readers get the most out of their technology.Prior to WIRED, Reece covered streaming at Business Insider.. Read MoreService WriterTopicsprivacyartificial intelligencepornsexOpenAIChatGPTchatbotsThe Search Engine for OnlyFans Models Who Look Like Your CrushPresearch’s “Doppelgänger” is trying to help people discover adult creators rather than use nonconsensual deepfakes.Jason ParhamHow to Organize Safely in the Age of SurveillanceFrom threat modeling to encrypted collaboration apps, we’ve collected experts’ tips and tools for safely and effectively building a group—even while being targeted and tracked by the powerful.Andy GreenbergWho’s Your Daddy? A ChatbotMore people are turning to AI to explore their BDSM fantasies, but some in the community feel it’s a cop-out.Jason ParhamThis AI Agent Is Designed to Not Go RogueThe new open source project IronCurtain uses a unique method to secure and constrain AI assistant agents before they flip your digital life upside down.Lily Hay NewmanGoogle Is Not Ruling Out Ads in GeminiWIRED spoke with Nick Fox, Google’s SVP of knowledge and information, about how AI is changing the company’s advertising business.Maxwell ZeffOpenAI Had Banned Military Use.The Pentagon Tested Its Models Through Microsoft AnywaySources allege the Defense Department experimented with Microsoft’s version of OpenAI technology before the ChatGPT-maker lifted its prohibition on military applications.Maxwell ZeffSears Exposed AI Chatbot Phone Calls and Text Chats to Anyone on the WebCustomer conversations with chatbots can include contact information and personal details that make it easier for scammers to launch phishing attacks and commit fraud.Matt BurgessSignal’s Creator Is Helping Encrypt Meta AIMoxie Marlinspike says the technology powering his encrypted AI chatbot, Confer, will be integrated into Meta AI. The move could help protect the AI conversations of millions of people.Matt BurgessThis AI Tool Will Tell You to Stop Slacking OffFomi watches you work, then scolds you when your attention wanders. It’s helpful, but there are privacy issues to consider.Justin PotHow to Hide Google’s AI Overviews From Your Search ResultsYou can avoid Google’s AI summaries in your search results by simply adjusting your query. Or just switch search engines altogether.Reece RogersHere’s What a Google Subpoena Response Looks Like, Courtesy of the Epstein FilesThe US Justice Department disclosures give fresh clues about how tech companies handle government inquiries about your data.Maddy VarnerGoogle’s AI Search Results Love to Refer You Back to GoogleThe company’s generative AI search tools increasingly cite its own services, like Google Search and YouTube, over third-party publishers.Reece RogersWired CouponsSquarespace Promo CodeSquarespace Promo Code: 20% Off Annual Acuity SubscriptionsLG Promo CodeLG Promo Code: 20% Off Your First OrderDell Coupon Code10% Off Dell Coupon Code for New CustomersSamsung Promo Code30% Samsung Coupon – Offer Program 2026Canon Promo Code10% Off Canon Promo Code + Up to 30% OffDoordash Promo Code50% Off Doordash Promo Code for New & Existing Users
Andrii Bidochko
CTO UBOS
Andrii Bidochko is an AI entrepreneur and researcher focused on AI agents, reinforcement learning, and autonomous systems. He writes about the technologies shaping the future of machine intelligence, from frontier models and agent architectures to real-world AI applications.