[ad_1]
Facebook and Instagram parent company Meta has said that it will add an invisible watermarking to the AI-generated content produced by its AI chatbot. The feature is a part of the company’s responsible development of AI tools, and safety and transparency efforts.
“Many images created with our tools indicate the use of AI to reduce the chances of people mistaking them for human-generated content.In the coming weeks, we’ll add invisible watermarking to theImagine with Meta AI experience for increased transparency and traceability,” Meta said.
The company also explained that the invisible watermark will be added with a deep learning model. Furthermore, it cannot be detected by the human eye but can be detected with a corresponding model.
It will be resilient to common image manipulations, which means that if there are any changes to an image such as cropping, colour change (brightness, contrast, etc.) and more, there won’t be any effect on the watermark.
Google, Microsoft use watermarking
The development comes months after the tech companies, including Meta, Google and Microsoft working in AI space were told to adhere to a certain set of rules for responsible development and implementation of the technology.
In August, Google launched a beta version of the SynthID tool for watermarking and identifying AI-generated images. Just like Meta’s tool, the technology embeds a digital watermark directly into the pixels of an image, making it imperceptible to the human eye, but detectable for identification. Similarly, in May, Microsoft pledged to introduce watermark for AI-generated images and videos.
Imagine image generator
Meta has also announced that it is expanding Imagine, its AI’s text-to-image generation feature, outside of chats, making it available in the US to start at imagine.meta.com.
“This standalone experience for creative hobbyists lets you create images with technology from Emu, our image foundation model. While our messaging experience is designed for more playful, back-and-forth interactions, you can now create free images on the web, too,” Meta announced.
“Many images created with our tools indicate the use of AI to reduce the chances of people mistaking them for human-generated content.In the coming weeks, we’ll add invisible watermarking to theImagine with Meta AI experience for increased transparency and traceability,” Meta said.
The company also explained that the invisible watermark will be added with a deep learning model. Furthermore, it cannot be detected by the human eye but can be detected with a corresponding model.
It will be resilient to common image manipulations, which means that if there are any changes to an image such as cropping, colour change (brightness, contrast, etc.) and more, there won’t be any effect on the watermark.
Google, Microsoft use watermarking
The development comes months after the tech companies, including Meta, Google and Microsoft working in AI space were told to adhere to a certain set of rules for responsible development and implementation of the technology.
In August, Google launched a beta version of the SynthID tool for watermarking and identifying AI-generated images. Just like Meta’s tool, the technology embeds a digital watermark directly into the pixels of an image, making it imperceptible to the human eye, but detectable for identification. Similarly, in May, Microsoft pledged to introduce watermark for AI-generated images and videos.
Imagine image generator
Meta has also announced that it is expanding Imagine, its AI’s text-to-image generation feature, outside of chats, making it available in the US to start at imagine.meta.com.
“This standalone experience for creative hobbyists lets you create images with technology from Emu, our image foundation model. While our messaging experience is designed for more playful, back-and-forth interactions, you can now create free images on the web, too,” Meta announced.
[ad_2]
Source link