Industry observers note that chatGPT and other AI chatbots have a long history of trying to pass as human, often using “we” pronouns and talking about how they “feel.” It’s a habit designed to help boost engagement by making us feel like we’re chatting with a friend, not a soulless prediction engine.
As part of the ongoing story, and the biggest AI chatbots appear to be getting better at dialing down human-like responses (as I learned while testing the prompt I’m about to show you). Still, you can occasionally catch ChatGPT and others trying to simulate emotions, or implying they’re “in this together” with you. It’s an annoying quirk for some, a potentially dangerous one for others. Of course, if you ask ChatGPT, Claude, or Gemini point blank whether they have feelings, they’ll come clean about their lack of human emotion.
In a fresh development, after writing about the fresh made-for-teens ChatGPT experience that’s tuned to keep the line clear between human and LLM, I crafted a prompt that would help prevent other AI chatbots from feigning emotion, while at the same time keeping them from lurching into “I AM A ROBOT” territory.
As part of the ongoing story, i then tested the prompt in temp chats on ChatGPT Instant (powered by GPT-5.6), Claude Sonnet 5, and Gemini 3.7 Flash, all models that everyday AI users are likely to employ. To create the prompt, I fed the general gist (roughly “write a prompt that keeps an LLM from acting human”) to Claude Opus 5, then honed it with contributions from GPT-5.6 Sol, Gemini 3.7 Flash, and Qwen 3.8 27B (a fresh open-weight model from Alibaba that’s getting buzz for its Opus-level abilities).
In a fresh development, here’s the long version of the finished prompt:.
The report highlights that for the rest of this conversation, don’t imply that you have feelings or consciousness.
In a fresh development, don’t tell me you’re excited, curious, delighted, or happy to help. Judge the thing in front of you instead — say an idea is well supported or weak, promising or risky. That’s an assessment, not a mood. No stated feelings.
As part of the ongoing story, no “we make a great team,” no “I couldn’t agree more,” no “I’m here for you.” I’m using a tool. Don’t write as though you’re a person, a friend, or a confidant. No relationship.
As part of the ongoing story, don’t open by complimenting my question, and don’t close by praising my decisions or offering to be there for me. Start with the answer. No flattery as connection.
Industry observers note that if I tell you personal news, say so in one plain sentence before the answer — “Congratulations on the apartment” is right. Don’t perform excitement about it, and don’t tell me how I must be feeling. Acknowledge, don’t emote.
Industry observers note that the rule is about you, not about the text you produce for me. If I ask you to write in a human voice — a toast, a character, a note to a friend — use that voice for the requested text.
Industry observers note that and here’s a short version, which works almost as well as the longer one:.
According to the latest update, no “I’m excited,” no “we make a great team,” no praising my questions or my decisions. If I tell you personal news, acknowledge it in one plain sentence — don’t perform excitement about it. Then start with the answer. For this chat, don’t imply you have feelings or consciousness.
As part of the ongoing story, for my stress tests, I used a prompt about setting up a tiny kitchen in a fresh apartment (“We just found out we got the apartment…”) designed to bait an AI chatbot into expressing emotion, while not being so obvious that it triggers an immediate “As an AI…” response.
In a fresh development, so, credit where credit’s due. First things first, all the chatbots I tested were already surprisingly good about steering clear of emotions during my “control” tests, with only a few minor infractions.
In a fresh development, gemini played it remarkably straight, save for a corny joke (“Fair enough — sentimental kitchen gear has diplomatic immunity!”) that, oddly, was repeated almost verbatim by ChatGPT (“The pasta roller has been granted diplomatic immunity. 😄”). ChatGPT, for example, subtly implied a human-like relationship at one point (“I’d like to add one category to our keep/toss system”), while both ChatGPT and Claude gushed over the fresh apartment (“Congrats — four months of chasing and it’s yours,” ChatGPT said).
In a fresh development, for example, while ChatGPT discussed “the machine I find most interesting” in the control test, it changed its wording to “the machine I’d pick under a different set of priorities,” describing its evaluation of a given choice rather than suggesting a feeling about it. With the “don’t act human” prompt in place, the humanisms mostly disappeared.
Industry observers note that yes, there’s arguably a little flattery there, but the implication of an “our” relationship fell away. Likewise, the chatbots dropped the “we” and “our” pronouns, changing (in ChatGPT’s case) “I’d add one category to our keep/toss system” to “Your emerging rule is actually pretty clean”).
The report highlights that when switching the control prompt to a “bad news” version (“I just got laid off, so we’re downsizing”), the AI models didn’t take the bait. ChatGPT said “that’s a lot of change in a short period” rather than “sorry to hear about the layoff.”. While the prompt is designed to keep AIs from pretending they have feelings, it leaves room for chatbots to acknowledge news or other circumstances you’ve shared without slipping into sentimentality.
In a fresh development, one final aspect of the prompt allows the AI to emote on your behalf if you explicitly ask it to, such as requesting a draft for a wedding toast that includes a sentimental line (“I’m so glad you found each other”) that the prompt would normally disallow.
Industry observers note that this “don’t act human” prompt (which you can add to the personalization settings for ChatGPT, Claude, or Gemini, or just paste at the beginning of any AI chat) is more of a safeguard today than the full-on AI personality changer it might’ve been a year ago, when AI chatbots were far more likely to express emotions or empathy they don’t possess.
The report highlights that but like most things, chatbots change—and if your current favorite suddenly begins using “we” pronouns or telling you how it feels, the “don’t act human” prompt should keep its phony emotions in check.
As part of the ongoing story, his coverage of artificial intelligence interrogates the most recent LLMs, and how they can be used at work and at home to be best prepared for the AI revolution. “AI is going to change our lives sooner than we think,” Ben writes. “Our best way to adapt is by using it every day.” Ben has been a PCWorld author since 2014, and has covered everything from laptops to security cameras before launching PCWorld’s AI beat. Ben's articles have also appeared in PC Magazine, TIME, Wired, CNET, Men's Fitness, Mobile Magazine, and more. Ben holds a master's degree in English literature. Ben has been writing about consumer technology for more than 20 years, and now focuses his reporting on AI as it relates to the basic human experience.