Uh oh! Turnitin is officially going to detect if you’ve used an AI chatbot in your essay

Guess it’s time to start your diss research x

Turnitin has officially developed software to detect if a student has used an AI chatbot, such as ChatGPT, in their essay. With the surge of assisted writing software, the plagiarism detective service has launched an AI Innovation Lab to work out if an essay has been written via an AI writing tool.

Turnitin has been working on AI-powered solutions “for several years” and it has since shared a preview into how its system can work out if an essay has been produced by a chatbot.

David Adamson, an AI scientist and former secondary school teacher, has offered an insight into how Turnitin can detect AI assisted essays. In the demo, David inserts an essay by ChatGPT with his own personal edits into Turnitin’s analysing tool and it figures out that half the essay was written by an AI. In fact, it found that 22 of the 43 sentences were all created by an AI tool.

Turnitin is hoping to reveal the full AI detecting tool soon but it is still currently in testing. Annie Chechitelli, Chief Product Officer at Turnitin, said: “Our model has been trained specifically on academic writing sourced from a comprehensive database, as opposed to solely publicly available content. As a result, Turnitin is more tuned to finding instances of potential dishonesty in student assignments.”

In a test to see what ChatGPT could really produce, recent graduate Pieter asked the AI to write an essay to see what grade he would get. In 20 minutes, the AI software produced the full 2,000 word essay on social policy. The essay got a score of 53 per cent which is the equivalent of a 2:2. The lecturer who marked the essay said the essay was “fishy” with no depth or analysis but it was reminiscent of work by “lazy” students.

The CEO of Turnitin recently said educators and students “have an opportunity and a responsibility to participate openly and honestly in the AI-assisted writing dialogue”, in order to “preserve integrity and trust in the education community.” He said: “We encourage you to have these discussions at your institution now and set achievable standards and expectations for your students around the acceptable use of AI-assisted writing tools.”

Universities across the country have recently made it an academic offence to use ChatGPT on assignments. Bournemouth University released a statement, telling students if all or part of an essay was detected as having been developed by an AI software, it would constitute as an academic offence.

The University of Edinburgh is currently reviewing ChatGPT’s “potential implications for assessments of students’ work and the effective of plagiarism detection software.” Queen Mary University of London recently told its students to “not ask an AI chatbot to write your assessments” as it would constitute as plagiarism.

Related stories recommended by this writer:

I got ChatGPT to write my essay to see if I could get a 2:1 from a Russell Group uni

Edinburgh University ‘reviewing’ how it detects plagiarism due to ChatGPT software

London universities warn against cheating on assessments with AI chatbots

Featured image credit via Adi Goldstein on Unsplash before edits.