You Should Still Check the “expert Advice” in Google’s AI Reviews for Credibility.
Google’s AI-powered search engine is receiving a series of updates designed to highlight first-hand information from forums and social media posts, as well as your trusted news sources and subscriptions. Specifically, AI responses will now show “expert advice” from online forums like Reddit, highlighting specific quotes and links to discussions related to your search queries. This “opinion preview” will also provide additional context for the discussion, including the author’s name, username, or community name.
It goes without saying that even when results are marketed as “expert advice” from forums you typically turn to for answers, you shouldn’t blindly accept them as gospel. As TechCrunch notes , AI is poor at recognizing sarcasm and humor, and Google’s AI reviews often recycle Reddit jokes and disguise them as serious advice. Generally, generative AI is also known for its tendency to hallucinate , simply making things up and confidently presenting anything from fake news reports to non-existent legal advice. While a New York Times analysis estimates that AI reviews are about 90% accurate, this means that at least one in ten responses will still contain errors.
In some ways, Google is making fact-checking easier with other AI-powered search updates: links to sources will appear in AI responses next to relevant text or list items, and hovering over embedded links will show a preview of the website before clicking. AI mode and AI reviews will also highlight content from your news subscriptions first, so you know the information is coming from sources you trust.
However, you should still make extra efforts to verify the credibility of the information provided by the AI. At a minimum, follow the link to the cited source to ensure that it actually says what the AI claims, and evaluate the source’s credibility. (Remember that “user-generated” doesn’t necessarily mean “expert opinion.”) And use indirect reading strategies to find authoritative sources that confirm or refute the AI’s claims.