Meta has announced that across its social media platforms artificial intelligence (AI) generated images must be labeled going forward.
The labeling system that will be titled “Imagined with AI” seeks to help users navigate the images that are created using AI software and those taken by a user traditionally.
Synthetic and non-synthetic content
Meta will be rolling out the watermark on the images created by its own generator and will be a standard for other major AI image generators such as Midjourney, Dall-E, and Bing Image Creator that have images uploaded to the company’s platforms.
The social media conglomerate is working with the Coalition for Content Provenance and Authenticity (C2PA) and International Press Telecommunications Council (IPTC) standards to clearly mark what “synthetic and non-synthetic content” is, according to the release.
C2PA is a Joint Development Foundation project from Adobe, Arm, Intel, Microsoft, and Truepic that hopes “to address the prevalence of misleading information online through the development of technical standards for certifying the source and history (or provenance) of media content,” according to the site’s mission statement.
Founded in 1965 in the United Kingdom, the IPTC is an established media and news agency coalition that develops and promotes industry standards.
Meta also said that audio and visual content are things they cannot identify from other companies that have been generated using AI. The company will add the option for users to tag this content as non-synthetic content to inform and protect users.
The company release said that users must use this “label tool when they post organic content with a photorealistic video or realistic-sounding audio that was digitally created or altered, and we may apply penalties if they fail to do so.”
The recent release strengthens Meta’s policies and rules on image and video content across all of its platforms, focusing especially on the well-being of minors.
Earlier this year, new content rules came into place to stop nude images from being sent to vulnerable users of Meta’s social sites as an optional feature.
With this momentum of policy and content changes, Meta is shaping up to be an industry leader in the rapidly evolving landscape of artificial intelligence and content verification.
The post “Meta makes AI-generated labels a standard across Facebook, Instagram and Threads” by Brian-Damien Morgan was published on 02/07/2024 by readwrite.com