Meta supercharges its AI: smart glasses get multimodal sight, new web image generator launches

Meta is turning AI generation into a group chat party trick.

Meta unveiled a massive rollout of updates across its AI ecosystem, introducing a new collaborative image-riffing feature called reimagine, launching a standalone web generator at imagine.meta.com, and bringing multimodal vision testing to its Ray-Ban smart glasses.

Riffing on images inside group chats

The headline social feature is reimagine, currently rolling out across Messenger and Instagram. When someone summons @MetaAI to generate an image in a group chat, any friend in the thread can press and hold that image, type a quick prompt, and watch Meta AI transform it into something entirely new.

Instead of image generation being a solo prompt-and-wait experience, it becomes a back-and-forth creative game. If you want a more traditional canvas, Meta also launched imagine.meta.com in the US. It is a standalone, free web tool powered by its Emu foundation model for standalone image creation.

To address provenance, Meta is deploying invisible watermarks to images created on the web tool. Applied via deep learning, these watermarks are invisible to human eyes but detectable by models even after screenshots, cropping, or color adjustments.

Smart glasses get eyes and apps get Reels

Hardware is getting the biggest brain transplant. On December 12, Meta kicked off an early access program for Ray-Ban Meta smart glasses owners in the US, enabling multimodal capabilities. The built-in camera can now inspect what you are seeing, allowing wearers to ask for photo captions on a hike or identify objects in hand, backed by real-time Bing search for sports scores and local navigation.

Across Facebook and Instagram, Meta AI is also creeping into core feeds. The assistant can now pull short-form Reels directly into chat threads when you ask for travel recommendations. Meta is also piloting AI-drafted birthday greetings, dating profile blurbs, and suggested DM replies tailored to creator tones.

Why it matters

Meta is playing to its unique strength: distribution and social graphs. While rivals like OpenAI and Anthropic focus on productivity workspaces, Meta is embedding AI into the connective tissue of daily messaging and everyday eyewear.

Turning image generation into a collaborative ping-pong match inside Messenger lowers the friction of using AI down to zero. You do not need to understand complex prompting when you are just riffing on a meme a friend made. For hardware, glasses with vision and real-time search make wearable AI feel significantly more practical than pure audio earbuds.

The obvious catch lies in platform fatigue. Injecting synthetic text into birthday greetings, group chats, and creator DMs risks flooding Meta platforms with hollow, algorithmic pleasantries. If conversations start feeling auto-generated, the social fabric Meta relies on could end up feeling remarkably synthetic.