The rise of AI-generated artwork and content online has become a growing concern for many in the tech industry, and even with people out of the tech business. Today, Meta announced that in a few weeks, it will begin labeling AI-based content on its Facebook, Instagram. and Threads social networks.
In a press release, Meta says these changes came from the recommendations of its internal Oversight Board, along with results from public surveys and advice from "academics, civil society organizations and others." Meta says it agreed with the Oversight Board that the older AI labeling rules were "too narrow."
Previously, the labeling for AI content was only for videos, video clips made by AI, or videos changed by AI tools to make it look like a person said something they didn"t actually say. As most of us know by now, AI tools like DALL-E, Microsoft Designer, and more can create or alter images as well as videos.
However, Meta said the Oversight Board recommended new ways to label AI content that did not violate the company"s Community Standards. Meta stated it has come up with new labels that will be placed on AI content starting in May. It added:
Our “Made with AI” labels on AI-generated video, audio and images will be based on our detection of industry-shared signals of AI images or people self-disclosing that they’re uploading AI-generated content. We already add “Imagined with AI” to photorealistic images created using our Meta AI feature.
Meta says it may put a more prominent "Made with AI" label on content if it determines that it may have a high degree of deceiving people. The company claims it also has almost 100 people assigned to find false or misleading AI-based content on its services. If a fact checker does find such content, it will be shown lower in a user"s Feed, along with an overlay that shows additional information.