This is a great example of where safety by design and having disclosure on how platforms operate are so important, because it is unlikely that we will be able to reliably identify whether an image is real or a deepfake; the computers will keep getting better and better at making these. That means we need to, instead, ask questions about whether we are weaponizing the platforms or making ourselves vulnerable to people abusing these images.
On December 12th, 2024. See this statement in context.