Meta is testing a feature that would place AI-generated content directly into Facebook and Instagram feeds. The material is designed around users' interests and current trends, making AI not just a tool people open, but something that can appear as part of the everyday social feed.
The test comes as Meta continues a broader push around Meta AI. The company says Meta AI is used by 400 million people monthly, with 185 million using it weekly, and it wants the assistant to become the most widely used globally by year's end.
How the feed test works
The new feature focuses on AI-generated images that adapt to each user's preferences. Instead of showing the same generated item to everyone, Meta says the content can shift according to what a user is likely to find relevant.
Users can also interact with the system through suggested prompts. Tapping those prompts guides the content in new directions, giving the user a way to shape what the AI produces next.
Swiping further can trigger more personalized content in real time. That detail matters because it turns the feed into a more active AI surface: the system is not only placing generated content in front of users, but also responding as they continue moving through the app.
One demonstration from Meta AI shows the image system asking users to imagine themselves as a video game character in space. The example points to a feed experience built around quick, personalized visual prompts rather than only posts from people, pages or accounts.
Meta AI is expanding beyond images
The feed experiment is part of a wider AI push across Meta's apps. Meta is also adding voice capabilities to its Meta AI assistant, allowing people to ask questions using voice rather than relying only on typed prompts.
The assistant is also gaining tools connected to photos. Users will be able to ask questions about photos and edit them, which places AI closer to everyday media sharing and discovery inside Meta's products.
Another addition is automatic translations of Reels videos. That expands the role of AI from generating new content to making existing video content easier to understand across languages.
Taken together, these features show a clear direction for Meta AI:
- AI-generated images inside Facebook and Instagram feeds
- Suggested prompts that let users steer generated content
- Real-time personalization as users keep swiping
- Voice features for the Meta AI assistant
- Photo questions, photo editing and automatic Reels translations
Why engagement is central to the test
As with other changes to its apps, Meta is likely trying to keep people on its platforms longer and more engaged. If AI-generated content adapts to each user and changes as they interact with it, the feed can become more responsive and harder to treat as static.
That is important for Facebook and Instagram because the feed is already a core place where attention is organized. Adding personalized AI content gives Meta another way to fill that space with material tailored to individual users.
The source article compares the direction, in an extreme way, to the Matrix movie, where machines feed artificial images to humans while using them for energy. The comparison is intentionally dramatic, but it captures the concern: AI could be used to create a stream of synthetic material designed mainly to keep people engaged.
The concern is not only that the content is generated. It is that the content is generated, personalized and refreshed in real time inside apps where people already spend attention. That combination raises questions about how much control users have over what they see and how much responsibility platforms take for the systems they build.
The rule-setting problem
The test highlights a basic tension in social media's next stage. Users need to decide how much time they want to spend on Facebook, Instagram and similar platforms. At the same time, the companies that run those platforms decide what kinds of content and engagement systems appear in the feed.
If AI-generated content becomes a normal part of social feeds, platform rules will matter. Meta can create tools that entertain, assist and personalize, but it also has to decide where the limits are and then stick to them.
The feature is still described as a test, but its direction is clear. Meta is exploring a feed where AI content is personalized to users, shaped by prompts and generated as people continue scrolling or swiping. That could make Meta AI more visible across Facebook and Instagram, while also making the debate over engagement, social effects and platform responsibility harder to avoid.