How to immediately cross-verify Google’s Gemini chatbot responses

This guide shows you how to double-check and cross-verify Google’s Gemini chatbot answers to ensure accuracy.

If you have a Google account, you are likely eligible to use the company’s generative AI chatbot Gemini much in the way you might use OpenAI’s ChatGPT - asking questions, learning new facts, or testing the limits of the new technology.

However, even advanced large language models from major companies such as Google are prone to making mistakes that nevertheless sound real and logical and look as if they are backed up with facts.

Users must remember that the technology underlying chatbots is still highly experimental and that using a chatbot-generated response for legal, academic, or journalistic purposes could lead to serious consequences.

However, Gemini can be useful in a pinch for everyday queries or low-stakes situations. Since Google is supported by a powerful search engine infrastructure, there is an easy way to cross-verify the responses you receive from Gemini, in just a few seconds.

(For top technology news of the day, subscribe to our tech newsletter Today’s Cache)

In our below example, you can see how a user asked Gemini whether pets are allowed on Air India flights.

Though the chatbot generated an answer citing Air India’s web page, an automatic Google search showed that some statements - such as pet carrier restrictions and limits on snub-nosed breeds - were not backed up by the needed evidence.