Meta, the parent company of Facebook and Instagram, announced that it will start operating Tag images created or manipulated by artificial intelligence (AI) on their platforms. This measure is part of Meta’s efforts to increase online transparency and prevent the spread of disinformation, especially ahead of the 2024 elections.
Meta already uses a visible watermark to mark images created by its own Imagine AI engine, which can produce realistic photos of people, places and objects. The company now plans to expand this practice to include images from third-party sources such as OpenAI, Google and Midjourney. Labels will indicate that the images were created by AI and provide more information about the source and process. The exact design of the labels is still being finalized but Meta shared a mockup of what they might look like, with the phrase “Artificial Intelligence Information” next to the images.
What about AI-generated audio and video? Meta admitted that it does not currently have the technology to identify such content on par with the images. However, he stated that the industry is working to build this capability and will update its policies accordingly.
Meta, meanwhile, will require users to indicate whether the audio or video content they upload is AI-generated or edited. If they fail to do so, sanctions will be imposed. Additionally, if the content is so realistic that it might mislead the viewer, Meta will place a more prominent tag with additional details.
Meta prepares to put an end to fake news
Meta also develops its own AI tools, such as the Imagine AI engine and Large Language Models (LLM). The company’s FAIR AI Research Lab is working on a new watermarking technique called Stable Signature that will make beacons more resistant to erasure. Meta also trains graduate students on its community standards so they can report content that violates its policies.
“ As the line between human and synthetic ingredients becomes increasingly blurred, people want to know where the line lies said Nick Clegg, the company’s head of international relations. “ People are often discovering AI-generated content for the first time, and our users have told us they appreciate the transparency of this new technology. That’s why it’s important that we help people know whether the photorealistic content they see was created using AI. “.
Meta said it will roll out tags for images created by artificial intelligence in the coming months. Many countries, including the USA, are preparing for the 2024 elections. The company hopes these labels will help users make informed decisions and prevent the spread of misinformation on its platforms.
However, Meta’s independently funded but independently managed Oversight Board recently asked the company to review its “inconsistent” manipulated media policy. After approving a resolution allowing an edited video of President Joe Biden to remain on Facebook.
