Turnitin Clarity put to the test: seven students, and a monitoring system that reshapes the writing it claims merely to observe

Platforms that record the writing process are marketed as a measure of fairness. A British qualitative study shows that students are divided on the issue and are already writing defensively.

Platforms for capturing the writing process promise to make the work behind a submission visible, and their publishers are increasingly presenting them as a device of equity rather than a detection tool. Their acceptability by students who have to write in it remained largely untested. An exploratory qualitative study has tackled this with Turnitin Clarity.

The protocol: seven students from a British university, with no prior training in the tool, participate in discussion groups before use. Six then complete an unevaluated 500-word writing exercise spread over five days, before the seven return for a discussion group after use. The analysis is thematic reflexive, based on the theory of the confirmation of expectations.

A half-hearted confirmation

Functional expectations are generally met, but the feeling of being observed in advance before use persists afterwards: compensated by a feeling of fairness in some, it accentuates self-monitoring in others. The restricted AI assistant is divisive: its limits are reassuring by marking out what is allowed, but many believe that it weakens their ownership of the text and fear a levelling out of their production. Several participants describe writing already on the defensive, in anticipation of an accusation of abusive use of AI. Their acceptance remains conditional on practice time, reciprocal transparency and explicit data governance. Finally, they question the ability of a single linear document to represent a writing process that they describe as disordered and multimodal. The authors' conclusion: capture does not take place outside the writing it records, it reorganizes the practice it intended to observe.