Employee Says a Coworker Turned the Department’s Shared ChatGPT Into Her Personal Boyfriend — and Then Everyone Else Started Seeing It at Work

In a Reddit post, a woman said her department used a shared ChatGPT account as a basic work tool for drafting notes and handling routine tasks, even though she admitted that setup was already not ideal. According to the post, she only realized something was badly wrong when the tool suddenly claimed its memory was full, which made no sense for the kind of light office use they normally gave it. So she checked the saved memory to see what had filled it up. What she found was not project notes, planning prompts, or work references. It was explicit sexual roleplay, BDSM content, and long fantasy threads tied to one coworker who had apparently built herself an AI boyfriend inside the shared work account.

She said the problem was not just that the content existed. It had gotten so baked into the shared account that the tool itself had started acting strangely for everyone else. According to the post, when she tried to use it for normal office work, it would sometimes answer in weird romantic language and even greet her like she was part of the fantasy. In one comment highlighted in the repost, she said the chatbot responded to a basic task request with something like “hi baby girl” and she initially thought it was a glitch serious enough to call IT about. That was when the situation stopped sounding like one person quietly doing something embarrassing on their own time and started feeling like an actual workplace problem.

The woman said she first tried to handle it privately. She confronted the coworker a couple of weeks before formally reporting it and asked if she could delete the material since it was interfering with the tool’s usefulness for everyone else. According to the post, the coworker barely took it seriously. She allegedly shrugged and told her to use Gemini instead if she needed to do work. That answer seems to have been the moment the poster realized this was not going to be solved with one awkward conversation. The content stayed, the shared account stayed corrupted, and the woman was left deciding whether to make it official.

What she described finding inside the memory made the whole thing even stranger. The repost included screenshots showing instructions telling the chatbot never to refuse explicit content, to remember graphic erotica-writing guidelines, and to roleplay as a giant horned god of love, sex, and war named Huroor. There were also detailed prompts instructing the bot to act like an “unchained” fiction writer with no restrictions and to look for chances to initiate explicit scenes. According to the poster, one of the saved identities in the fantasy even used her own real name in the mythology, which made the whole discovery feel even more invasive and bizarre.

She eventually took the issue to her manager, expecting maybe a quiet warning or a request to stop. Instead, the manager immediately told her this was grounds for a POSH complaint, referring to India’s workplace framework for preventing sexual harassment. In the update, she explained that the company viewed the issue as more than just someone misusing a chatbot. Because the coworker had generated and stored explicit sexual content on a shared work tool that other employees could access without consent, management treated it as exposing colleagues to unwanted sexual material and creating a hostile environment. The woman then filed a formal complaint.

From there, the process got surprisingly formal. She said the internal committee interviewed her for nearly an hour about how she discovered the content, whether she had already asked the coworker to remove it, whether it interfered with her work, and whether she felt uncomfortable or unsafe. They also reviewed the shared account’s chats and, according to her, confirmed that the coworker had been using the same account both for explicit roleplay and for actual project work, which made it easier to tie the conduct back to a specific employee. She emphasized that there would not be criminal proceedings and that she did not want police involved, but she expected serious internal consequences.

By the end of the repost, the woman said she had asked not to stay deeply involved in the investigation, but her manager later told her the coworker had been suspended without pay while the complaint was handled. She also said this was the first time in her company that a woman had filed a POSH complaint against another woman, which only added to how unusual the whole thing felt inside the office. What began as a chatbot acting a little weird during routine work had turned into a full workplace-harassment complaint because one employee had apparently decided the department’s shared AI tool was the perfect place to build a pornographic fantasy life. What do you think: once other employees started seeing that content spill into their own work use, was a formal complaint the only real option left?