Meta plans to add watermarks to content generated by Imagine with Meta for increased transparency and traceability. The invisible watermarks, which will be generated with an AI model and detectable with a corresponding model, are resilient to common image manipulations. This move comes as tech firms face increasing pressure to make it clearer that works were generated by AI, particularly in light of the rise of Deepfakes and AI-generated inappropriate images.
Key takeaways:
- Meta has launched a new standalone generative AI experience, Imagine with Meta, that allows users to create images by describing them in natural language. The tool is powered by Meta’s existing Emu image generation model and is free to use for users in the U.S.
- The company plans to add invisible watermarks to content generated by Imagine with Meta in the coming weeks for increased transparency and traceability. The watermarks will be resilient to common image manipulations.
- Watermarking techniques for generative art are not new and have been adopted by other tech firms like Microsoft and Google. The pressure is increasing on tech firms to make it clearer that works were generated by AI.
- Regulations are being put in place requiring that generative AI vendors mark generated content. China’s Cyberspace Administration has issued such regulations and in the U.S., Senator Kyrsten Sinema emphasized the need for transparency in generative AI, including by using watermarks.