Is my friend using ChatGPT to text me?
Your friend usually texts "lol ok". Today you got three warm, balanced paragraphs. Hmm.
Seven giveaways
- A sudden change of voice: longer, tidier, kinder, more "thoughtful" than usual.
- Perfect punctuation from someone who never uses capital letters.
- Em dashes — from someone who has never typed one in their life.
- Balanced feelings: "I completely understand where you're coming from, and I also feel…"
- Summaries: they repeat what you said back to you before answering.
- No in-jokes, no nicknames, no references to last weekend.
- Slow then perfect: a long pause, then a long, flawless message.
A playful way to check
Ask something only they would know: "what was the name of that awful café?" A chatbot can't help with that, so the answer comes back in their usual voice, typos and all. Or just ask, nicely: "did chatgpt write that? 😂"
Is it actually bad?
Not always. People use AI to word hard messages (apologies, breakups, condolences) because they care and are worried about getting it wrong. If the message matters, the feeling behind it can still be real. It's worth a gentle conversation rather than an accusation.
The giveaways above are the ones you'll practise in the daily game: two real people, one AI, four questions.
Questions
Can I prove they used ChatGPT?
No, and detectors are unreliable. Treat the signs as a reason to ask, not proof.
Why does AI texting feel off?
Because it is smooth in a generic way. Friends text with shared history, shortcuts and mistakes, which chatbots leave out.