← all tags
// tag: ai-detection
does chatgpt watermark your text now under the eu ai act
Not yet, and only in the EU. What OpenAI's textGrain actually turns on, when it lands, how well it detects — and why nobody outside OpenAI can check it.
does removing claude's ai watermark actually work
Short answer: not in any way you can verify. Why the removers strip the wrong thing, what actually breaks the mark, and the detector nobody can access.
granta asked claude if the story was written by claude. claude was very helpful.
A purpose-built detector flagged the story at 100% confidence. Granta's response: ask the AI that may have written it. Claude issued a non-verdict. Granta published.
