Claude users are rage-quitting because Anthropic tagged their ‘original’ work
It seems Anthropic decided that your intellectual property belongs in a digital lineup. By embedding invisible, tamper-proof watermarks into every word Claude spits out, the company has effectively turned every power user into a walking, talking AI-labeled exhibit.
The new watermarking system, rolled out in early August, doesn't mess with your font or style. Instead, it manipulates the statistical choice of words during generation to create a unique, invisible fingerprint that survives copy-pasting and even manual editing. This isn't just some random experiment—it's Anthropic's way of playing nice with the EU AI Act, which demands that AI-generated content be machine-identifiable.
For the professionals, consultants, and students relying on the tool, this feels like an unwanted stamp of mediocrity on their work. The core issue is that the watermark persists even after heavy human intervention, meaning the tool essentially claims partial ownership over the final draft. Consequently, a small but vocal group of users is jumping ship to competitors like Cursor or Grok, hoping to find a cleaner, unmarked slate for their projects.
While Anthropic insists these marks are purely for identification and don't signal full AI authorship, the optics are disastrous for anyone trying to pass off AI-assisted work as purely human labor. Even though the company claims no massive spike in cancellations, the message is clear: users don't want their silent partner leaving digital breadcrumbs all over their professional reputations. In a world obsessed with authenticity, being caught with an invisible metadata tattoo is apparently the ultimate tech-bro nightmare.
Source: Business Insider
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.