SocialMinds - A Social Media Marketing Podcast · SocialChain

SOCIAL IN SIX [102]

March 16, 2026·17 min·1 clip
Meta is testing an AI chatbot that uses your personal data to recommend products, signaling a shift in how we shop online.
March 16th opens the roundup. Taylor Mitchell joins the host for her Social Minds podcast debut, and the episode moves quickly from platform news into what marketers might have to deal with next. Meta is testing chatbot shopping. The reported Meta AI feature would answer shopping questions with product carousels that show images, prices, brand details, and external retail links. Familiar territory. The host ties it to ChatGPT tests in the US with Sephora, Target, and Wayfair, treating AI chat as part search engine and part shopping assistant. Then Instagram gets trickier. A test shows shopping overlays on Reels, and creator Julia Berolzheimer only finds out after followers message her about products she never tagged or approved. That is the uncomfortable part. Instagram says Meta's system pulls from business catalogues, and Meta says it is not taking commission during the test. Still, trust gets messy fast. Audiences may assume the creator endorsed the products, even if the creator had no idea they were there. Brand deals make it worse. If creators have exclusivity clauses, automated suggestions could put the wrong products in front of followers. X's chatbot comes back into the conversation too. After earlier scrutiny around Spicy Mode, X quietly adds an opt-out toggle for photo modification and generation. That small setting carries weight. The segment also covers Unhinged Mode, where The Athletic reported offensive responses after football fans asked the chatbot to roast rivals. Near the end, Taylor talks about keeping historic data visible alongside new formats, so teams can still compare performance as platforms change. The practical lesson lands cleanly: social teams need to watch where automation starts blurring consent, endorsement, and measurement.
Listen to the show on