Social media platforms Facebook and Instagram, both owned by Meta, have recently publicized their plan to combat the rise of deepfakes and misinformation, especially with the approaching elections. They will be implementing ‘Made with AI’ labels to AI-generated content, marking a significant shift in their policy.
Meta’s new policy will start coming into action from May. Facebook and Instagram will begin labeling AI-generated videos, images, and audio content with ‘Made with AI’. This is a significant expansion from their previous policy, which only addressed a small portion of manipulated videos.
The Vice President of Content Policy, Monika Bickert, stated in a recent blog post that the policy’s scope is much broader now. The company will also apply separate and more noticeable labels to digitally altered media that could potentially deceive the public on important matters. These labels will be applied irrespective of whether the content was created using AI or any other tool.
Meta’s revised approach signifies a change in how they handle manipulated content. Instead of removing a limited number of posts as they did earlier, they will now keep the content up but provide viewers with information about how it was made.
Meta had previously announced a plan to detect images created using AI tools from other companies. They planned to use invisible markers built into the files. However, they did not specify when they would start implementing this plan.
A spokesperson from Meta informed Reuters that the new labeling approach would apply to content posted across Facebook, Instagram, and Threads services. However, WhatsApp and Quest virtual reality headsets will be governed by different rules.
Meta will start applying the more prominent ‘high-risk’ labels immediately, added the spokesperson. These labels will be used for media that poses a significant risk of deceiving the public, especially on important matters.
This new policy is being introduced just a few months before the presidential elections in November. Tech researchers have warned that these elections may witness a transformation due to the rise of generative AI technologies.
Political campaigns have already started using AI tools in places like Indonesia, pushing the boundaries of the guidelines provided by Meta and generative AI market leader OpenAI.
In February, Meta’s oversight board criticized the company’s existing rules on manipulated media as ‘incoherent’. This was after reviewing a video of President Biden, which had been altered to suggest inappropriate behavior. The board suggested that the policy should also apply to non-AI content, which can be just as misleading as content generated by AI.
The rise of AI and deepfakes puts a responsibility on social media platforms to protect their users from deceptive content. The new labeling initiative by Facebook and Instagram is a step in this direction. However, it remains to be seen how effective this approach will be in the upcoming elections.
