Feb 09, Kathmandu- Meta will label images uploaded to Facebook, Instagram and Threads. The company is going to make this kind of arrangement targeting the upcoming elections around the world.
In addition to this, it is said that the company will also punish users who do not reveal the reality-based video or audio parts created by AI. Meta's global affairs president Nick Clegg said in an interview that AI-generated media is becoming increasingly difficult to distinguish from reality, and he is taking steps to deal with it.
The US government is also asking technology companies to make arrangements for watermarking content created by AI so that it can be easily identified. According to Clegg, Meta is building a tool to identify content created by AI. Also, Meta has already applied the watermark of 'Imagine with AI' to the pictures prepared by its Imagine AI generator.
While the company may not be able to identify everything on its own, he said Meta is on "high alert" as to how this type of media content could be used to manipulate it.
Meta is also working with groups like the Partnership on AI, which is working on a campaign to identify the authenticity of existing content. Adobe recently released the Content Credential System, which includes information about the origin of content in image metadata.
Similarly, Google has extended its SynthID watermark for audio files after releasing the beta version for images.