Google Gemini's latest feature will verify the accuracy of AI responses
Now, users can take control and verify the authenticity of the content generated by the chatbot.
Big language models like Gemini can sometimes get things wrong. In fact, many companies in the generative AI space caution users about potential inaccuracies or "hallucinations" in AI-generated content, something I can bet many of us have experienced first-hand.
Now, Google is addressing these concerns letting users take control and verify the authenticity of the content generated by the chatbot, by introducing a "double-check response" button to enhance reliability.
The way this works is that, whenever you use Gemini for a query, you will find this button below AI-generated text, allowing you to verify the accuracy of information using Google Search. This feature is available on both the app and the web version of Gemini AI, by the way.
To enhance clarity, statements are classified into three categories for cross-verification. Text highlighted in green indicates that Google found similar content online, and there's even a link so you can check it out for yourself. Orange highlights indicate no relevant findings, while unhighlighted text suggests there just isn't a whole lot online to compare the info to.
According to Google, "The double-check responses feature helps you assess the credibility of Gemini’s statements using Google Search to find content that’s likely similar or different."
This feature came shortly after Google began rolling out AI Overviews, a feature powered by the Gemini foundation model that presents AI-generated answers and summaries to some users' search queries, appearing above the traditional search results.
The whole point of this is to make sure you can trust what Gemini is telling you by double-checking, and feel more confident about the information you're getting.
This is a big win for the over 2 billion users using Google Gemini according to Google's CEO Sundar Pichai, and a significant step towards enhancing the reliability and trustworthiness of AI-generated content. So, the next time you chat with Gemini, you can be sure you're getting the straight scoop.
Henry Chikwem
