Detecting deepfakes involves a combination of visual inspection, audio analysis, and leveraging AI-powered tools. Start by examining the video or audio for inconsistencies, such as unnatural eye movements, inconsistent lighting, or lip-syncing issues. You can also use online tools and software that utilize machine learning algorithms to identify deepfakes. Additionally, fact-checking the content and verifying its source can help determine its authenticity.
Author here. I expect most to be ironed out too, regardless of whether they're publicized or not. The engineers working on these models are aware of the problems