Facebook Adds AI Mode That Reads Your Public Posts

Meta began rolling out a wave of new AI features on Facebook on June 15, led by 'AI Mode' — a search tool that answers plain-language questions by pulling together public posts from across the platform, including Groups and Reels. Alongside it come photo and video editing tools: collage cutouts, montage transitions, AI presets that swap your outfit, hairstyle or accessories, and a sports-themed 'Wear It' option that drops your favorite team's jersey onto your profile picture.

AI Mode is Meta's answer to the way people increasingly start a search by typing a full question instead of a few keywords. Ask it something and it synthesizes a reply from public discussions happening in Groups and short-form Reels, rather than sending you off to a list of links. It is the same pattern Google has been pushing with its own AI Mode, and it lands after a steady drip of Meta AI launches this year — animated profile pictures in February, an AI Marketplace assistant in March, a creator assistant earlier in June, and a Reddit-style Forum app with an 'Ask' tab in May.

Why this matters to you

Two things. First, your public Facebook posts — and the public posts in any Group you have ever written in — are now answer fodder for a chatbot that strangers can query. 'Public' on Facebook has always been loosely understood, and AI Mode makes the consequences concrete: a years-old comment can resurface inside someone else's answer with your name still attached to it. Second, the answers are only as reliable as the unvetted posts behind them. Meta's own framing acknowledges the risk that outdated or misleading user content gets laundered into a confident-sounding summary — the same failure mode that has repeatedly embarrassed Google's AI Mode.

The editing tools are the crowd-pleasers, but AI Mode is the real shift: Meta is turning a social feed into a question-answering surface, which keeps you inside the app instead of leaving for a search engine. My read is that accuracy is the thing to watch — a summary that sounds authoritative but quietly mixes a 2019 rumor with a 2026 fact is worse than no answer at all, because it is exactly the kind of reply that stops you from checking.