Turnitin recently announced they have developed a new functionality relating to the use of AI tools (such as Chat GPT) in written assessments.  Turnitin required universities to take a very quick decision on the use of the functionality, as they announced it would be used from early April for all Turnitin customers unless they opted out.  This note explains the BU decision not to use the  functionality at this point in time.

What is the new functionality?

It uses an AI detection model to produce a report showing the overall percentage of the document that may have been generated by AI tools.  This functionality is an addition to the Similarity Report which is already available to BU staff when marking assignment submissions in Turnitin.  Where a university chooses to receive the AI detection report, this would be visible to staff alongside the Similarity Report (but opting out of the new functionality will not change current access to/use of Turnitin Similarity Reports and marking tools ).

Why has BU decided not to use the AI detection functionality?

We have decided that introducing the functionality at this time would not be fair or appropriate for either staff or students.  Turnitin has introduced the functionality on very short notice, with limited information and without the opportunity for preview or testing by individual universities.  We need to take longer to assess the functionality and ensure that any use at BU is within an appropriate framework of policy, guidance and communication.  Many other UK universities have made the same decision to opt out of this functionality at the current time.

Our decision is due to a number of factors, in particular:

• We would be introducing a new assessment tool mid-academic year while entering an intensive assessment period. It would not be appropriate to do this without planned, structured support and guidance in place for staff and very clear advance communications with students.
• The very short lead-in time from when Turnitin announced the new functionality to the date when it became active means there has been no opportunity for us to preview or test out the functionality, to consult with staff about the guidance and support they would need, or to communicate sufficiently with students.
• We also need more time to consider how the Turnitin functionality would fit within BU policy on academic offences. We are currently considering changes to this policy to cover use of AI tools, in the context of our current approach to assessment.  At the same time, we are conscious that our position on AI also requires further consideration and development over the longer term, taking account of views around appropriate student use of AI tools and the impact of the availability of AI tools on assessment design.
• Turnitin have not provided any detailed guidance on how the AI detection tool works or evidence/validation of the reliability of its accuracy.
• Whilst academic integrity is a key concern, making these new reports visible to staff without providing clear policy and guidance on how to respond to the information would place staff in a very difficult position and is also likely to cause concern to students. Taking the time for a more considered approach will enable us to robustly assess the new Turnitin functionality tool in the context of our developing policy on AI, determine whether it is something we should adopt in the future, and provide appropriate support and guidance for staff and students in advance.

Further policy enhancements on the use of AI in assessment will be communicated with staff in the coming weeks.

Further support and guidance

• For advice on BU assessment and academic offences policy please contact Academic Quality
• For advice interpreting the Turnitin Similarity reports contact your Faculty Learning Technologist.