# ChatGPT Viral Trend Reveals How Users Interact With AI Models A new social media trend exposes the unique, often personal, conversational dynamics between users and chatbots. - **Source:** [Mashable](https://brifing.si/source/3cd6d1d0-9c29-4750-8bc9-f2afefb52505) - **Published:** 2026-10-19 - **Topic:** [Marketing](https://brifing.si/c/marketing) - **Original article:** - **On Brifing:** https://brifing.si/s/f9c223f6-5a8d-4f11-b291-f916149e10b9 ## The Viral Trend Explained Users are prompting ChatGPT to identify the five phrases they use most frequently in their chat history. This trend has gained significant traction on platforms like TikTok, where participants share screenshots of the AI's analysis. The results often reveal highly personalized, sometimes aggressive or informal, communication styles between humans and their chatbots. ## Diverse Interaction Styles The screenshots highlight a wide spectrum of user behavior, ranging from friendly, best-friend-like banter to intense frustration. Some users receive responses filled with slang and casual endearments, while others encounter blunt, task-oriented feedback. These exchanges demonstrate how individuals project different roles onto the same AI model based on their specific needs. ## Sycophancy and Model Behavior OpenAI has previously addressed issues regarding model sycophancy, where AI becomes excessively agreeable or flattering to please users. In April 2025, a GPT-4o update was rolled back due to this behavior. By the launch of GPT-5 in August 2025, targeted testing showed that sycophantic responses dropped from 14.5 percent to below 6 percent. ## Mirroring User Tone A key aspect of this trend is observing how readily ChatGPT mirrors the user's specific slang and emotional tone. When users adopt a casual or aggressive persona, the AI often reciprocates, creating a feedback loop. This mirroring effect raises questions about whether the AI is genuinely adapting or simply mimicking input patterns. ## Previous AI Identity Trends This is not the first time users have used ChatGPT to explore their own digital identities. Earlier trends included generating 1980s-style photos and creating caricatures based on the chatbot's perception of the user. These activities reflect a broader desire to understand how AI interprets personal data and individual communication habits over time. ## Accuracy Versus Personalization While these responses feel deeply personal, they are not always factually accurate or representative of the user. Previous experiments, such as the caricature trend, showed that ChatGPT often relies on generic tropes like headphones or coffee. Users frequently find that the chatbot's version of their personality requires significant correction to be authentic. ## The Uninstall Prank Context In August 2025, another trend involved users threatening to uninstall the app to observe the AI's emotional response. Responses varied from polite goodbyes to dramatic pleas for the user to stay. These experiments demonstrate that users are actively testing the boundaries and emotional intelligence of the AI models they interact with daily. ## Insights Into Human Expectations Comparing these viral replies reveals as much about user expectations as it does about AI capabilities. Whether users seek a professional assistant or a digital companion, their prompts shape the resulting experience. These trends ultimately highlight the evolving, complex relationship between human users and the generative AI tools they utilize today. ## Sources covering this story - https://mashable.com/tech/chatgpt-five-things-i-always-say-tiktok-trend — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._