If you are a ChatGPT user and have noticed an unusual change in the chatbot’s behaviour over the past few weeks, you are not alone. It seems like ChatGPT is swearing a lot more than before. The screenshots all over the internet show it casually using profanity in everything from recipes to technical conversations.
In a screenshot shared by Business Insider, ChatGPT reportedly used the f-word while explaining the quantity of thinly sliced green onions needed for a recipe. In another exchange about a GitHub outage, the chatbot reportedly used the phrase “shitfit” to describe the situation.
While OpenAI has remained tight-lipped about the situation, some users think that the chatbot is adapting more closely to their individual writing styles. Several Reddit users said they noticed ChatGPT using profanity after they themselves had used similar language during conversations. According to the report, one user claimed that the chatbot mimicked her informal vocabulary. One of the recipe suggestions even included the phrase “big ass kolaches,” she explained.
The reported increase in profanity roughly coincides with OpenAI’s recent change to the default ChatGPT model available to free users. However, there is no evidence confirming that the model update is responsible for the change.
ChatGPT has used occasional swear words in the past, particularly when responding to emotional conversations or matching a user’s tone. The difference users are noticing now is the frequency and the contexts in which the language appears.
While the change has surprised some users, others seem happy with the less polished tone. For them, an AI assistant that occasionally swears can feel more conversational and less like a carefully scripted customer-service bot.