People are pasting suspicious texts into AI chatbots to check them

More people are pasting odd text messages into AI chatbots to ask one question: is this a scam? A review on tech-insider.org compares three of the best-known options, ChatGPT, Claude and Gemini, on how well they spot fake texts.

What the review covers

The review looks at ChatGPT (GPT-6), Claude (Claude 5.5) and Gemini (Gemini 4 Argon). The idea is simple. You copy a suspicious message, paste it into the chatbot and ask whether it looks like a scam. The full article text was not available for this story, so the details of its tests and results are not covered here.

Why people are doing this

Irish regulators say fake delivery, bank and tax refund texts are flooding phones this year. These messages often look routine, which makes them easy to trust in a busy moment. A chatbot can feel like a quick way to get a second opinion.

What is not known yet

No independent lab has published a head-to-head scam detection score for the three chatbots. That means there is no neutral number showing which one catches the most scams, or how often any of them gets it wrong. The material also does not say how the review itself measured results.

What to do

• Use an AI check as a second opinion, not a final answer.
• Never tap the link in a suspicious text, even to test it.
• Report real scams to the authorities.