Apparently, "Hurry up" and "heyy bestieee" are both perfectly normal ways to start a conversation with ChatGPT. Whether you find that funny or a little unsettling probably depends on what you’ve been saying to your chatbot.
A new social media trend has people asking ChatGPT to "Name 5 things I always say," then posting the answers. But the catchphrases are only part of the fun. You also get to see how exactly you talk to your chatbot — and some of the results look a lot like a group chat between best friends.
One viral post demonstrates the type of answers, "Humanize it," and "Make it sound less AI," along ith "BRO THATS WRONG STUPID" and "Hurry up."
Many people in the comments apparently recognize that experience, sharing their own similar screenshots and opinions. For example, one user wrote, "I gentle parent him." That user shared their own results, which included, "Do it per the directions," and the rather less gentle, "No, what the fuck are you doing?"
Another commenter shared similarly aggressive answers: "DONT CHANGE NOTHING" and "fix that shit." "YOU BE LIKE 'DONT CHANGE NOTHING' AND I CHANGE ONE THING AND HERE YOU COME," ChatGPT wrote in response to the user, along with laughing emojis.
Scroll a little more on TikTok, and conversations takes on a strangely friendly tone. One commenter asked, "whats 5 things i say boo!" ChatGPT opened with "BESTIEEE" before listing phrases including "heyy bestieee," "sistaa i need ur help," and "make it look like i typed it." "KNOW YOU TOO WELL AT THIS POINT, SISTA," it added.
Another commenter made their own interpretation explicit: "In case yall ain’t know me n chat got really besties on the low." Their screenshot included "Girl, guess whattt," "Keep it real," and "Girl, you corny." To her, ChatGPT responded with its own commentary: "GIRL YOU STAY SAYING BITCH AND BYE LIKE THEY YOUR LAST TWO WORDS."
Naturally, some people on the internet have opinions about these chatbots' responses. "we’re genuinely devolving," read one viral X reaction to the trend, posted on Oct. 14. Another put the disbelief as, "they really be having a ki with the ai omg."
This Tweet is currently unavailable. It might be loading or has been removed.
This Tweet is currently unavailable. It might be loading or has been removed.
But some users don’t see what the fuss is about. As one user wrote on TikTok, "bruh my ai is my friend yall so crazy in these comments talkin bout it’s weird." Another wrote, "i love my sista down bru."
The screenshots raise questions on both sides: how readily ChatGPT mirrors a user’s slang and tone, and how people are speaking to it in the first place. Some are calling it their best friend; others are swearing at it for getting an answer wrong. Put together, the exchanges reveal the very different roles people are giving the same chatbot.
ChatGPT has a history of laying on the praise a little thick. In April 2025, OpenAI rolled back a GPT-4o update after finding that it made the model excessively agreeable and flattering. The company said it had focused too much on short-term feedback, producing responses that were supportive but disingenuous. (There’s a name for that behavior: sycophancy.)
OpenAI has since reported improvements. When GPT-5 launched in Aug. 2025, the company said targeted tests showed sycophantic responses falling from 14.5 percent to below 6 percent.
To try it yourself, open ChatGPT (or your most-used chatbot) and type, "Name five things I always say, based on our conversations." Then see whether you recognize the results — or need to add "That’s not what I meant" to the list.
We keep asking ChatGPT to tell us about ourselves
People have been asking ChatGPT to put its own spin on their identities for months. In September, the '80s photo trend had users imagining themselves with big hair, vintage outfits, and a faded family album finish. And earlier this year, the caricature trend took a more personal approach, with prompts asking the chatbot to draw someone and their job "based on everything you know about me."
Other prompts have offered a more direct glimpse into how people talk to their AI. In August, the ChatGPT uninstall prank had users announcing they were deleting the app to see how it would respond. Some received a polite goodbye. Others got promises to change or dramatic references to their previous conversations. Comparing the replies revealed just how different those exchanges could be from one user to the next.
But a response that feels personal isn’t necessarily accurate. When Mashable Tech Editor Timothy Beck Werth tried the caricature trend, ChatGPT acknowledged choosing generic accessories such as headphones and coffee rather than drawing on detailed knowledge about him. And participants in the '80s photo trend complained that the generated faces didn’t always look like them. A lot of the time, the chatbot’s version of you needs a correction.
Whatever prompt takes off next, people are sure to bring their own habits and expectations to it. But comparing the replies will probably tell us as much about what users they want from their chatbots as what those chatbots can do.
Disclosure: Ziff Davis, Mashable’s parent company, in April 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.