Meta’s new AI tool can put you into any video with just your selfie

Meta has introduced a new artificial intelligence tool called Emu Video that enables users to insert themselves into AI-generated video clips using only a selfie and a text prompt.

The announcement marks Meta’s latest move in its broader strategy to integrate generative AI into consumer-facing applications, with implications for content creation, marketing, and social media engagement.

Using multimodal AI capabilities, Emu Video allows users to describe a setting - such as “a desert at sunset” or “a cyberpunk cityscape”- and upload a single photo of themselves. The system then produces a short, photorealistic video that places the user within the generated environment, maintaining consistent perspective, lighting, and facial alignment.

How does Emu Video work?

Emu Video builds on recent advances in AI that combine natural language processing with image and video generation. According to Meta’s engineering blog, the tool is designed to handle three modalities: text for scene generation, images for subject insertion, and video for final output synthesis.

The model leverages Meta’s previous work in image generation (Emu) and animation (Make-A-Video), integrating them into a unified pipeline that can produce high-quality clips in seconds.

Meta says Emu Video is part of its larger goal of developing more immersive and expressive AI-powered tools for its suite of platforms, including Instagram, Facebook, and the Horizon metaverse.

Applications and limitations

The tool is expected to appeal to creators, influencers, and digital marketers looking to produce visually striking content quickly and without the need for advanced technical skills or expensive production equipment.

However, its release has also raised concerns about potential misuse, particularly around deepfakes and unauthorised video manipulation.

Meta stated that Emu Video will include several guardrails, including:

• Visible watermarking to indicate AI-generated content
• Limits on video duration and use cases
• Monitoring tools to detect and flag potentially harmful content

Emu Video is currently being tested internally and with select creators, with a wider release expected in the coming months. Meta plans to refine the tool based on user feedback and expand its use cases to include collaborative projects, storytelling, and short-form entertainment.