Does Turnitin detect Claude?
Yes
Yes. The detector is model-agnostic. It scores the statistical shape of the writing, so Claude output is flagged the same way GPT output is.
How it works
Detectors of this class do not maintain a list of known model outputs. They measure how predictable each token is given the ones before it. All instruction-tuned assistants produce unusually low-variance prose, which is the signal being measured.
The part people miss
Because the signal is 'unusually smooth writing' rather than 'was generated by model X', clean, formal human writing lands in the same region.
Flagged for something you wrote yourself?
There is a right order to handle it in, and editing the document first destroys your best evidence.