Generative Artificial Intelligence

Generative AI is a new and exciting/terrifying development. It holds great potential but carries significant risks to academic integrity and traditional working practices. You can find out more about the University's official position regarding generative AI, and read our guidance to students in the use of generative AI in their learning and assessment.

Read our top tips for helping students to understand the role of generative AI in learning and assessment and to make the most of the power of generative AI, whilst mitigating against the risks.

Tips for using generative AI

Set clear boundaries

Be clear from the outset about what uses of Gen AI are permitted on your course and at University more generally. Point students to the University guidance on the use of generative AI in their learning and assessment.

Develop and promote AI literacy

A good place to start would be the University's AI Fundamentals online course, which gives a basic introduction to what we mean by "AI", a guide to the ethical considerations around AI use, a guide to some of the tools available and how to use them, and a summary of the University's guidance on the use of AI. In addition to this, the library have created a LinkedIn Learning pathway called A beginner's guide to generative AI which is a good first introduction, both to AI itself and to the some of the many courses available through LinkedIn. These courses range from basic introductions to deep dives into specific topics.

Encourage your students to develop their understanding of generative AI so that they can use it to help rather than hinder their education. The University has published some advice for students on using AI in their learning and assessment which gives a brief explanation of what generative AI is and explains how it can be used appropriately in a higher education context.

Encourage students to have a critical approach to GenAI

Generative AI can create large amounts of credible-looking content very quickly, but it also frequently hallucinates, providing false information, including fake academic references.

To help students develop a critical approach to Gen AI, one option is to design activities that require them to critically engage with its output. This will allow students to gain an understanding of how artificial intelligence differs from human intelligence, of the kinds of errors that Gen AI tends to make, and the importance of developing their own academic skills. Critiquing AI output will also help them develop those skills.

It may also help to demonstrate the limits of Gen AI by showing examples of the kinds of tasks it routinely fails, or by asking the tool to set out its own limitations or to give examples of which elements of its response it is most unsure of.

Augmentation over automation

Show students how to use AI to enhance their learning rather than undermine it. Students can use AI to bypass their learning activities by automating certain tasks such as researching and writing their assignments. But they can also use AI to do better, deeper research and to develop their written work by providing feedback on their plans and drafts and suggesting improvements.

You can tailor this advice to your module or programme by looking at generative AI use through the lens of your learning objectives. For each objective, explain what kind of generative AI use (if any) would make it easier to achieve, and what would make it harder.

Design assessments with Gen AI in mind

There is no such thing as a Gen AI-proof assignment, but it's possible to design tasks that are harder for Gen AI to do well. Examples include:

Presentations: either live in person, or recorded via video and or audio.
Process-based assignments such as keeping a reflective journal, or a scaffolded assignment
In-person written assessments, such as exams.
Assignments based on the students' own research or lived experience.

Another approach is to integrate generative AI into the assessment (e.g. with a generative AI evaluation project). You can find ideas for assessments that incorporate generative AI in the assessment menu from UCL which is part of their guidance on designing assessments for an AI-enabled world.

Guard your data

Whatever tool you or your students may be using, always be aware that any data you enter will be transmitted outside the institution and will often be used to train the AI model. Do not enter any personal, confidential or restricted data into an AI tool. For more information about information security considerations, please see our AI usage guidance (information security).