Even today there are many open models that generate images without any labeling. These AI models can be shared and improved indefinitely, even if most major AI players take a multifaceted approach to labeling AI content. The generative genie is out of the bottle and there is no way to put it back in.
Since there will always be unlabeled AI content, you can’t count on watermarks to tell you what’s true. As AI tagging becomes more common, people may think that an image or video that lacks one is automatically legitimate. It’s a dangerous but understandable mistake for someone to make. While the existence of invisible watermarks may give people some sense of security as they explore an increasingly distorted Internet, they are not a reliable way of knowing what is true.
“The future we live in will increasingly be a world of unlimited content,” Rose said. “It’s economics 101, supply and demand. Supply is through the roof and the value of content will continue to fall. But increasingly, there will be a demand for authenticated content, a way to prove that something is what it claims to be.”
The scale of AI content alone will make it impossible to label everything. Google’s 100 billion SynthID images and videos are just the tip of an iceberg that has already eclipsed authentic content on the Internet. It may be wiser to focus on verifying the truth rather than revealing falsehoods.
“There is a limit to the amount of content that can be created with real physical cameras in the world,” explained veteran photojournalist and fellow Starling Mike Caronna. “There is no limit to the amount of synthetic content that can be created. That’s why we want to focus on what is scarce, and what is scarce is truthful, high-value information. If we can secure that and prove that it is authentic, we can defend ourselves against the liar’s dividend.”
