Meta, the parent company of Facebook and Instagram, has unveiled its strategy to tackle the issue of AI-generated content by implementing “AI generated” labels on its platforms. This move is a proactive measure to combat the spread of false information in the lead-up to the 2024 elections and beyond.
Nick Clegg, President of Meta Global Affairs, shared the company’s plans in a recent blog post. The aim is to provide users with a clear distinction between images created using artificial intelligence and those captured conventionally.
This decision comes at a time when concerns have been mounting within the tech industry, as well as among lawmakers and information experts, about the rising influence of AI tools capable of producing remarkably realistic visuals. The combination of these tools with the rapid dissemination of content on social media platforms poses a substantial risk of disseminating misleading information, particularly during elections in various countries, including the United States.
Meta’s labeling initiative will encompass images generated using a variety of third-party AI tools, including those from Google, Microsoft, OpenAI, Adobe, Midjourney, and Shutterstock. Building on the success of their existing “imagined with AI” label for photorealistic images created with their in-house AI generator tool, Meta’s new labels will now be applied to images produced with third-party AI solutions.
To ensure consistent and accurate labeling, Meta is collaborating with leading AI tool developers to establish common technical standards. These standards will involve incorporating invisible metadata or watermarks within AI-generated images, allowing Meta’s systems to identify them with precision.
These “AI generated” labels will be introduced across Facebook, Instagram, and Threads, and will be available in multiple languages to ensure widespread accessibility.
Clegg highlighted the importance of transparency when it comes to AI-generated content. He noted that users often encounter AI-generated content for the first time and appreciate knowing when such technology is involved. This labeling approach will be maintained throughout the coming year, which encompasses critical elections worldwide, enabling Meta to gain insights into user preferences and the evolving landscape of AI-generated content.
However, it’s important to note that the industry-standard markers used to label AI-generated images will not yet extend to videos and audio generated by artificial intelligence. To address this limitation, Meta plans to introduce a feature that will allow users to identify and disclose AI-generated video or audio content when sharing it. Users failing to disclose such content may face penalties.
Meta’s commitment to transparency also extends to safeguarding against users removing the invisible watermarks embedded in AI-generated images, given the increasingly adversarial nature of AI-generated content creation.
In related news, Meta has also announced an expansion of its anti-sextortion tool, “Take it Down,” in collaboration with the National Center for Missing & Exploited Children. This tool empowers users, especially teens and parents, to create unique identifiers for intimate images, enabling platforms like Meta to promptly identify and remove such content.
Initially launched in English and Spanish last year, “Take it Down” will now be accessible in 25 languages and additional countries. This expansion follows recent Senate hearings where Meta CEO Mark Zuckerberg and other social media leaders faced scrutiny regarding their platforms’ protections for young users.
As Meta takes these measures to address the challenges posed by AI-generated content and protect the integrity of its platforms, the tech industry continues to grapple with the evolving landscape of information dissemination and digital manipulation in an era where visual authenticity is increasingly difficult to discern.