Meta, the parent company of Facebook and Instagram, has recently unveiled significant revisions to its policies concerning digitally manipulated media, just ahead of impending elections that will serve as a litmus test for its capacity to combat deceptive content originating from artificial intelligence technologies.
In an announcement made on Friday, the social media behemoth disclosed its plan to introduce “Made with AI” labels starting in May for AI-generated content such as videos, images, and audio shared on both Facebook and Instagram. This initiative marks an expansion of their previous policy, which had predominantly targeted a limited scope of altered videos. Monika Bickert, the vice president of content policy, elaborated on these developments in a blog post.
Furthermore, Meta intends to implement distinct and more conspicuous labels for digitally modified media that presents a “heightened risk of materially misleading the public on significant matters,” irrespective of whether the content was crafted using AI or other editing tools. The rollout of these enhanced “high-risk” labels is set to commence immediately, as confirmed by a company spokesperson.
This strategic shift signifies a departure from the company’s prior approach to manipulated content, transitioning from mere removal of select posts to a more transparent strategy of leaving the content accessible while furnishing viewers with insights into its creation process.
Previously, Meta had introduced a mechanism to identify images generated using third-party generative AI tools through embedded invisible markers within the files, although an official start date was not specified at the time of the announcement.
According to a company representative, the labeling framework will be applicable to content shared on Facebook, Instagram, and Threads, while distinct guidelines govern its other platforms such as WhatsApp and Quest virtual-reality headsets.
These policy adjustments precede the upcoming US presidential election in November, a pivotal event that experts caution could be significantly impacted by the proliferation of generative AI technologies. Political campaigns have already started leveraging AI tools in regions like Indonesia, pushing the boundaries of existing guidelines established by Meta and leading generative AI provider OpenAI.
In a noteworthy development, Meta’s oversight board criticized the company’s current regulations on manipulated media as “inconsistent” following a review of a video featuring Joe Biden that was circulated on Facebook last year. This video had been altered to falsely insinuate inappropriate behavior by the US president.
Despite concerns raised by the oversight board, the video remained accessible, as Meta’s prevailing policy on “manipulated media” prohibits misleadingly edited videos only if they are AI-generated or if they fabricate statements attributed to individuals. The board emphasized the necessity for this policy to encompass non-AI content, which can be equally misleading, along with audio-only content and videos depicting individuals engaging in actions or making statements they never actually did.
