How to Tell If You're Chatting With a Bot Online

The short answer: a chat message cannot prove who wrote it

A person can write a polished answer in seconds. A bot can use slang, make typos, remember details, and wait before replying. One odd response—or one convincing response—does not prove whether a person or software is on the other end. That is especially true in a brief stranger chat, where you have little context.

A NIST study of AI-written summaries found that detection performance varied substantially across the systems tested. That research was not a test of your particular chat partner. Treat online 'AI detector' scores and viral trick prompts as guesses, not identity verification. The more useful question is whether the service tells you when you are talking to AI and whether the conversation respects your boundaries.

Check the platform label before judging the writing style

Look at the chat header, matching screen, and any introductory notice. Some services openly provide AI companions or automated support; others offer only person-to-person matching. A clearly labeled AI conversation is a choice you can make. An account pretending to be a person while concealing automation is a different problem.

On ChatUNO, the AI companion is identified in the chat interface, while Text and Voice matching are options for meeting another person. If your goal is a human conversation and the companion opens, check the label and use Next to search for a person again. Do not infer identity from the pace of the typing indicator alone.

• Read the name and any AI or automated-chat label in the header.
• Check whether the platform describes the conversation as AI or person-to-person.
• Use the available leave or next control when the type of chat is not what you wanted.

Notice patterns, but do not turn clues into a verdict

Repeatedly ignoring a specific question, giving a generic answer that could fit anyone, or changing basic facts can justify slowing down. So can replies that steer every topic to the same link, product, or off-platform account. These are reasons to protect your time and information—not proof of a bot. A distracted, shy, translated, or scripted human conversation can look similar.

The reverse is also true. A funny local reference, a typo, a voice note, or a photo is not proof that the speaker is who they claim to be. People can use AI to draft a message, and media can be copied, prerecorded, or generated. Judge the interaction by its transparency and behavior over time rather than a single 'gotcha' test.

Ask directly, then decide what you need

If it matters to you, a simple question is better than a puzzle: 'Is this an AI chat or am I talking to another person?' A trustworthy platform should answer through its labeling. Another participant may answer honestly, refuse, joke, or lie; their reply alone is not independent verification.

You can follow up on something they actually said: 'You mentioned that game—what do you like about it?' This helps you see whether the conversation is enjoyable, not whether a hidden detector would classify the reply as human. Do not demand a selfie, ID document, live camera call, or exact location as a test. Real people may have good reasons to decline, and giving a stranger more private data creates new risk.

• 'I prefer chatting with a person. Is this an AI conversation?'
• 'I am not comfortable moving to another app yet. Can we stay here?'
• 'This is not the kind of chat I wanted, so I am going to leave.'

Separate bot suspicion from scam warning signs

A disclosed AI assistant is not automatically a scam, and a human can be a scammer. The Federal Trade Commission warns about fake profiles that build trust and then ask new online friends for money. It also advises slowing down when someone creates urgency or asks for personal information. Those behaviors matter whether the messages were typed by a person, generated by software, or both.

End the exchange if someone asks for gift-card codes, a transfer, cryptocurrency, a verification code, or a login link. Do not send money to an online contact you have not independently verified. Avoid installing a 'proof you're human' app or running commands supplied in chat. If you already paid or shared a password, contact the payment provider or secure the account promptly through official channels.

What to do when a conversation feels deceptive

Stop adding personal details. Keep a record of the account name and the messages that explain your concern, without continuing the exchange just to gather proof. Use the platform's block and report controls for impersonation, unwanted pressure, suspicious links, or financial requests. You can also simply leave a conversation that does not feel right; you do not need to win an argument about whether the other side is a bot.

If the issue is identity deception rather than automation, our related guide to catfishing explains how to evaluate a pattern of claims without becoming invasive. If money is involved, use the money-request response guide below. The safest next step is usually to slow down, not to perform more elaborate tests inside the same chat.

Choose the kind of conversation you actually want

There is nothing wrong with choosing a disclosed AI companion for a casual exchange or choosing a person-to-person chat instead. What matters is knowing which experience you are entering and having an easy way to leave. A good chat service should make that distinction clear enough for you to decide.

If you want to meet a person on ChatUNO, try Text or consent-based Voice mode. Keep the first conversation light, do not share identifying details just to prove you are real, and use Next, End, Block, or Report whenever you need them. A comfortable conversation should not require you to surrender your boundaries.