Meta's AI Labelling For Images Has Gone Totally Wrong: Here's What Happened
Meta is tracking AI content but other images are also getting the tag. Meta AI is slowly making its way to different platforms and more users but the company faces a major issue with its technology. Reports say that Meta AI is labelling real images with the Made by AI label which suggests Meta has some bigger issues to tackle as basic aspects of the AI tool is falling short.
The Problem with Meta's AI Labelling
Meta is supposed to add these labels for actual images that are created using AI. But it seems the platform is marking real photos that have been edited using AI tools as Made by AI as well, which isn’t how it should work. The report from TechCrunch quotes photographers who have seen Meta make this mistake and the worrying part is that the company has not exactly given the reason for these labels on real images. However, the report has Meta quoted, saying that it is evaluating its approach for using these AI labels and how it is able to determine which image is eligible to come with the badge.
The Need for Accuracy
Nobody is doubting the need for companies like Meta to keep a check on the quality and presence of AI content on its platforms but wrongful labelling shows that the actual mechanism to detect these images is severely lacking. The report also says that Meta wants to fix this issue by working with companies that offer AI-based tools for editing and creation. These changes need to happen soon as the US is headed into General Elections later this year, and the last thing Meta wants is disinformation about AI content when they are genuine and credible.
The Challenges Ahead
The timing is also not apt for the company as it pushes Meta AI into its platforms like WhatsApp, Instagram and Facebook, which has a reach of over billions of users with countries like India also now able to access these tools.
