Meta Responds to Google Gemini with ‘Meta for AI-Generated Images’

In response to Google’s Gemini AI model, Meta, the parent company of Facebook, has launched Imagine with Meta, a standalone generative AI experience accessible on the web. 

Similar to OpenAI’s DALL-E, Imagine with Meta uses the company’s Emu image generation model to produce high-resolution images from text prompts. The tool is currently free for users in the US, generating four images per prompt.

Meta aims to enhance transparency by introducing invisible watermarking to the Imagine with Meta AI experience in the coming weeks. This watermark, applied through a deep learning model, is designed to be imperceptible to the human eye but detectable with a corresponding model. It will withstand common image manipulations like cropping, colour change, screenshots, and more.

The move aligns with the broader industry trend, as companies like Microsoft, Google, Shutterstock, and others adopt AI-based watermarking standards. China’s Cyberspace Administration has mandated similar regulations, emphasizing the need for transparency in generative AI. 

This also aligns with the global push to clearly identify content generated by AI, especially in the wake of concerns about deepfakes and AI-generated abusive images.

Meta’s Imagine with Meta, positioned for creative hobbyists, is set to provide users with a unique text-to-image generation experience beyond messaging platforms.