Meta launches AI Mode search on Facebook, sourcing answers from public posts across platforms

Meta is launching AI Mode as a new search option on Facebook, powered by its Muse Spark AI model. The feature generates answers by pulling from publicly-posted content across Meta's platforms including Facebook, Instagram, and Threads.

How AI Mode works

When users search on Facebook, AI Mode now appears alongside existing search modes like "People" and "Marketplace." According to Meta, the feature "uses Meta AI to give you answers grounded in what people are saying publicly across our apps" rather than displaying traditional link-based results.

Users can ask follow-up questions in response to the AI-generated results. Meta says Muse Spark will "over time unlock new features that cite recommendations and content people share across Instagram, Facebook, and Threads."

Additional AI features

AI Mode is part of a broader rollout of AI features starting June 15, 2026, including:

• Photo presets that digitally swap sports jerseys onto fans
• Suggestions for collage templates
• Integration with Meta's Reddit-like Forum app, which includes similar AI search capabilities

Context: The platform content race

Google has previously drawn criticism for surfacing Reddit threads in both traditional search results and AI Overviews. Meta's approach mirrors this strategy by leveraging its own vast repository of user-generated content across its social platforms.

The feature raises questions about how publicly-posted content is used to train and inform AI systems, particularly as users may not have explicitly consented to their posts being used as source material for AI-generated answers.

Technical details

Pricing, model specifications, and performance benchmarks for Muse Spark have not been disclosed. Meta has not confirmed whether Muse Spark is a new model or a variant of its existing Llama architecture.

What this means

Meta is positioning its massive archive of public social media posts as a competitive advantage in AI search, potentially challenging Google's dominance. The company now has a closed-loop system where user-generated content feeds its AI products, which in turn keep users within Meta's ecosystem. This could accelerate the shift from traditional link-based search to AI-generated answers, though accuracy and attribution remain open questions when sourcing from unverified social media posts.