Cambridge Rejects Turnitin Terms Allowing Student Work for AI Training

Summary

The University of Cambridge refused to sign a new Turnitin licence agreement after proposed changes to the end-user terms appeared to allow student submissions to be used in developing or training AI tools. Cambridge Students’ Union said the university is operating under a modified arrangement and will continue using the previous terms until July 2027, meaning student work submitted during the current academic year will not be used for AI training under that arrangement. The university is exploring alternatives to Turnitin, while the Students’ Union supports the decision. The university said it acted after identifying possible data-privacy and intellectual-property implications, and that Turnitin agreed to postpone issuing a new EULA until September 2027. It also said Turnitin had confirmed that no student or staff data had been used to develop generative AI models or large language models and that it had no plans to use student essays in that way. Turnitin said it does not train a generative AI model and pointed to its Trust Centre and privacy information. The company had proposed changes that institutions would have needed to accept to continue using the service, which uses machine learning to compare submissions with previous work and academic sources. Usage varies across Cambridge: some departments check all assessed online work, while others conduct spot checks or investigate only suspicious submissions. The dispute has spread beyond Cambridge, with student unions at several UK universities raising concerns. Southampton said it would immediately begin phasing out Turnitin and would not renew after the 2026-27 academic year; York said it would review the terms. The article reports that Turnitin products are used by 98% of UK universities, according to Uni Compare.