Artificial Intelligence

A Working Group has been set up to consider the University's approach to the use of Generative Artificial Intelligence Tools in teaching and learning settings. We will take a positive, educational approach to the potential use of generative AI tools as part of teaching, learning and assessment at Reading. 

The Position Statement on Generative AI is the first output from the Working Group. Ahead of the start of the Autumn Term the Working Group will deliver an Annex to the Academic Integrity and Academic Misconduct policy to support staff and students. A user guide for students is in development by colleagues from Study Advice and should be available early in the new term. 

On a longer term basis the Working Group will be developing a suite of guidance materials alongside an over-arching policy on the use of generative AI tools to support staff and students. Materials will be added to this website in due course.   

In the interim please feel free to contact the Co-Chairs of the Working Group (Professor Alan Howard and Dr Alexandra Hayward), or its Secretary (Richard Sandford), if you have any queries. 

Policy and Guidance

The University Board for Teaching, Learning and the Student Experience has recently approved amendments to Section 9 of the Assessment Handbook (Academic Integrity and Academic Misconduct) to make it clear that the misuse of Generative AI Tools will normally be considered an act of plagiarism.

Additionally, an Annex to Section 9 has been produced on Generative Artificial Intelligence Tools, Academic Integrity and Academic Misconduct which sets expectations for Schools and students. The Annex also gives an indication of what might be considered misuse of Generative AI Tools.

Further support materials are in development and will be shared with colleagues when they are ready.

Generative AI Detection Tools

With the nature of AI-generated content changing rapidly, caution should be applied to the use of any ‘ AI detector’. For example, Open AI (the creators of ChatGPT) launched a tool to distinguish between AI-written and human-written text but withdrew it due its low rate of accuracy. (source https://openai.com/blog/new-ai-classifier-for-indicating-ai-written-text).

Turnitin released AIWDC (Artificial Intelligence Writing Detection Capabilities) functionality in April 2023. We have opted not to switch this on at the University at this point. We need to have confidence in its reliability and effectiveness, understand its limitations, and ensure there is policy and advice around its use if we are to switch it on. Early use in the sector has indicated that while Turnitin AIWDC is reasonably able to detect simple copy-pastes from basic AI prompts, this ability drops when common methods of deceiving AI generation are applied to these outputs. There is also concern around “false positives” where a student’s work is incorrectly identified as AI-generated.

We will keep a watching brief on these detection technologies, and consider if they have a role as part of a wider approach to AI. Turnitin themselves advise that teachers should not rely on detectors and “will need to apply their professional judgment, knowledge of your students, and the specific context surrounding the assignment" when considering if AI has been used. (source https://www.turnitin.com/blog/understanding-false-positives-within-our-ai-writing-detection-capabilities)

Useful Resources

Key Documents on the Use of Generative AI in HE - Details will follow shortly.

Using Generative AI in Assessment Design - Details will follow shortly.