Instagram is now automatically generating headlines and descriptions for users’ posts that show up in Google Search results, frequently without users’ awareness and occasionally with misleading or incorrect framing, as reported by 404 Media. This issue was highlighted when author Jeff VanderMeer observed an Instagram post of his—a video featuring a bunny munching on a banana—appearing in Google search results with a clickbait-style title he did not create. Other posts, including one from a library in Massachusetts, displayed unrelated promotional titles.

Over the years, Google has been incorporating more social media content into its search results. Beginning July 10, Instagram made public posts from professional accounts searchable on Google.

404 Media verified that this issue impacts various users. For example, a cosplayer’s video showed up in search results with a title about finding cosplay spots in Seattle, which the user did not compose. Google informed 404 Media that it does not create these headlines, but instead extracts text directly from Instagram.

Meta later confirmed that Instagram recently started utilizing AI to craft titles for posts displayed in search engines, intending to assist users in better comprehending content, albeit recognizing inaccuracies. The company pointed users to a Help Center page that explains how to opt out of search engine indexing.

Experts quoted by 404 Media assert that these headlines are created for search bots and incorporated into the page’s title tags, meaning typical visitors do not see them on Instagram. Users and creators have criticized this approach for stripping posts of context, misrepresenting their work, and leading to low-quality, spam-like search results.

This reporting brings to light wider concerns regarding how AI-generated metadata could misrepresent online content and affect the reputations of creators.