The internet’s best-selling author is about to get the recognition it deserves.
What happened: Anthropic is rolling out a new invisible watermark that will be embedded in any text, image, or file generated by its AI models that were released after August 2, making it far easier for online platforms to detect and flag content generated by Claude. The company says it's working on adding the function to its older models, as well.
Why it’s happening: Whether it's social media posts, articles, or music, companies have realized how fed up people are with AI slop. Spotify announced yesterday that it would start labelling AI artists and blocking them from its recommended playlists, while Substack launched a tool last month that scans posts to flag if they are AI-generated.
Why it matters: The dead internet theory — the once fringe prediction that the entire online ecosystem would be overrun with bots — is eerily close to our reality today. Over half of new articles posted on the internet are AI-generated, nearly 35% of all internet traffic is now from bots, and a quarter of social media posts over 250 words are written by AI.
While detection tools like Anthropic’s won’t stop AI content altogether, they could help filter the slop from the feeds of those who would like to avoid it.
Our take: We’ve become accustomed to lawyers citing non-existent AI legal cases in open court, consultants generating reports with fake experts, and even politicians accidentally reading their own chatbot prompts out loud in the provincial legislature. We’re going to find out pretty quickly who still remembers how to do their jobs without Claude.—LA
