Instagram is facing criticism after reports emerged that the platform has been quietly attaching AI generated headlines and descriptions to user posts without their direct knowledge. The issue came to light when 404 Media revealed that Instagram pages now contain sensational or search optimised titles embedded within their code. These titles are not visible within the app itself, but can appear within external search engine results. Soon after the report, an Engadget editor noticed similar behaviour on their own posts. According to early findings, these automatically generated text snippets appear to be designed to boost Instagram’s visibility on Google by enhancing how individual posts are indexed, but many users are raising concerns about the accuracy, tone, and transparency of the content being attached to their work.
Examples of these unintended SEO style captions show how the automated system can misrepresent user generated content. A post by Engadget’s Sam Chapman regarding a board game he developed had an entirely inaccurate description assigned to it. Instead of referencing his actual board game Bloomhunter, the AI generated text described a different puzzle game entirely. This mismatch suggests the system is pulling generic or unrelated information to create descriptions. Additionally, 404 Media highlighted an instance where author Jeff VanderMeer uploaded a simple, untitled clip of a bunny eating a banana. The platform attached an embellished headline that sounded more like a promotional article than a caption for a brief video. A library in Massachusetts promoting a reading of VanderMeer’s work received similarly inflated wording, suggesting that this auto generated system is being widely applied across different account types.
Some creators have expressed discomfort with the stylistic and factual inaccuracies of the generated content. Multiple cosplayers, including Brian Dang, reported finding unexpected and overly generic titles linked to their posts. Dang noted that the tone of the AI generated text did not align with how he represents his work and suspected the titles were created at scale through a language model. For creators who rely heavily on authenticity and personal branding, the idea that their content may be reframed without warning raises concerns about misrepresentation. The issue becomes more complicated when considering that these AI generated titles exist only in the site’s metadata, meaning creators may not notice the changes unless they check their posts through external tools or search results.
Technical analysis confirmed that these AI generated titles and descriptions appear within the tags and other metadata fields of Instagram’s web pages. Users can detect them through tools like Google’s Rich Result Test, which reveals how search engines process and display the content. These titles differ from Instagram’s auto generated alt text, which is meant to support users with low vision. Instead, they function as SEO elements, crafted to help search engines interpret and rank content. Meta responded to inquiries by stating that it recently began using AI to create titles for posts intended for search engine indexing, claiming the feature is meant to help people understand shared content more easily. The company also noted that users have the option to disable indexing, though doing so removes the post entirely from search discovery, making it harder for accounts to be found. Meta acknowledged that the auto generated text may not always be completely accurate, reinforcing the concerns of users whose posts have been inaccurately described.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
