Using AI in education

What is generative AI?

Generative AI (GenAI) is a type of AI that creates new text, images or other content through tools such as ChatGPT, Gemini, Claude and Copilot. For example, a GenAI tool could be asked to explain attachment theory, and it will provide a well-written, coherent answer. 

Text-based GenAI uses large language models (LLMs) trained on vast collections of books, articles and websites. These models learn patterns in language and generate responses by predicting the next letter or word in a sequence. As they rely on historical data, their responses reflect both the strengths and weaknesses of that material.

What is The Tavistock and Portman’s position on students’ use of GenAI?

We want you to explore GenAI responsibly and use it appropriately and ethically. We want you to develop the knowledge and skills to be able to use GenAI effectively to support your learning and help you succeed in the future. 

This guidance is also applicable to staff involved in teaching or supporting learning at The Tavistock and Portman.

What are the limitations of GenAI?

Although GenAI can produce fluent and convincing text, it has limitations:

• it may present information that is out of date or unreliable, since it learns from past internet-based material
• it can provide incorrect information and citations, sometimes referred to as ‘hallucinations’
• responses reflect biases present in the original training data
• it can struggle with nuanced, contextual, or contemporary issues, where little accurate information is available

Errors can be subtle and hard to detect, so you should never assume outputs are correct without checking against reliable sources.

Ethical concerns

The use of GenAI also raises wider ethical questions which you should consider before using it:

• environmental impact – training and running large models requires significant energy, contributing to climate change
• human labour – AI systems rely on people to check and refine outputs during training, which can cause psychological strain and exploitation
• bias and fairness – AI can reinforce biased decision-making if not carefully monitored
• copyright – training data often includes copyrighted material, raising questions about ownership and fair use
• privacy – extensive use of personal data in training and operation can lead to privacy violations

Using AI on your course

GenAI can be used as part of your studies at The Tavistock and Portman, but you must use it appropriately. Read our guidance on using AI appropriately before using it.

GenAI tools

There are many GenAI tools to choose from. There are general, well-known GenAI tools such as Copilot Chat and ChatGPT, which can be used for multiple purposes. There are specialist tools designed for academic tasks such as Research Rabbit, and there are AI features embedded into existing systems.

About our guidance

Our guidance is adapted from the University of Essex’s Artificial Intelligence Guidance for Students.