Skip to main content Skip to navigation

Artificial Intelligence

This page was updated on 9/3/2026 to provide further clarity of the History Department's position on GenAI.

This page explains the History Department's approach to Generative Artificial Intelligence (GenAI). This guidance is designed to help you to understand when it is acceptable (or not) to use GenAI in your history assignments. If you have further questions, you can speak with your Personal Tutor, Seminar Tutor, or Year Director. Please note, however, that as GenAI continues to evolve, and historians begin to integrate it into their practice, what is acceptable use of GenAI may well change. Therefore please check this page regularly for updates.

As this page explains, even when GenAI use is acceptable, it may not be desirable. It is often better to learn to do something yourself before delegating it to a machine, because this enables you to better judge what GenAI is telling you.

Key Principles

  1. You must be able to demonstrate intellectual ownership of your own work. This means that you must not use GenAI to replace your own intellectual labour, which includes thinking, reading, and writing.
  2. You must not use GenAI tools to gain an unfair advantage or to create content that is presented as your own work. This is Academic Misconduct.
  3. You must not use GenAI to complete any part of your assessments that you would not ask another human to do for you. GenAI must not be used to complete any element of your work that is assessed under the department marking criteria.
  4. You must acknowledge GenAI use in the same circumstances, and in the same way, as you acknowledge any other sources or tools that you use for your assessments.
  5. You should be aware of the data protection implications of GenAI and you must respect the intellectual property rights of others. This means only using university-licensed GenAI tools (currently Microsoft CoPilot) if you wish to input course materials or documents (including essay questions, lecture slides, articles) that you have not authored or which include others’ data (e.g. oral history interviews). You are welcome to use other GenAI tools for your own work, but be aware that your intellectual property may not be protected when you do so.
  6. If an individual assignment has specific instructions regarding the use of GenAI that differ from those in this handbook, you should follow the instructions for that assignment.
  7. If in doubt, don't use GenAI for an assignment task without checking with your seminar tutor first.

Key Terms

  • Automation: Getting machines to do things that people once did. Includes mental as well as physical automation. For example, a pocket calculator automates mental arithmetic. Artificial intelligence (AI) is a more complex kind of mental automation.
  • Non-Generative AI: A type of artificial intelligence designed to optimise processes, analyse data, and make predictions. Basic spelling-checkers and grammar-checkers are examples of non-generative AI tools.
  • Generative AI (GenAI): A type of artificial intelligence that can create new content, such as text, images, audio, or video, by learning patterns from existing data. ChatGPT, CoPilot, Claude, and Gemini are examples of GenAI tools.
  • Large language model (LLM): A type of GenAI specifically trained on large-scale textual data to understand and generate human-like language through statistical pattern recognition.
  • Prompt: The input or question a user gives to a GenAI model, which the model then uses as a basis to produce a response.
  • Hallucination (in AI): When GenAI produces content that sounds plausible but is factually incorrect or entirely made up.

GenAI and the Study of History

Historians take a deliberative approach to studying the past, critically assessing their sources as well as scholarship in the discipline. In a history degree you learn a wide range of skills, including critical thinking, analysis, synthesis, a sensitivity to the strangeness and unknowable nature of the past, attentiveness to complexity and contingency, the deployment of evidence, and the development of your individual voice. As you develop these skills over the course of your degree, it is vital that you maintain, and can demonstrate, intellectual ownership of your work.

Some of these skills can be plausibly imitated by GenAI, such as summarising large amounts of text (synthesis) or “analysing” texts. However, there are ethical, emotional and embodied dimensions to historical study which cannot be replaced by GenAI. GenAI can help historians with some tasks, such as finding patterns in large amounts of data, but other historical tasks are done poorly (or cannot be done at all) by GenAI, such as conducting interviews, or handling original archival documents.

Historians are especially sensitive to some of the pitfalls of GenAI tools, many of which are built into their design. For example, GenAI tools:

  • Are trained on a specific corpus of texts which (though large) leaves out a vast number of records available to a trained historian (e.g. those in physical archives);
  • Are not trained to recognise change over time, or to take account of the date of the publication of a source when assessing its usefulness or accuracy, so they can provide out-of-date information;
  • Are not trained to consider the origin of the texts they are drawing on, so they draw no distinction between peer-reviewed scholarship and anything posted on the internet;
  • They can sometimes hallucinate: they can simply invent things, whether footnotes, quotations, or “facts”;
  • They can be vague about their sources, in part because they synthesise many different sources;
  • They tend to flatten out individuality and creativity, producing generic prose and predictable ideas.

GenAI has immense potential in historical work. But it should be used with great caution, especially for history, where the processes of reading, thinking, and writing are so intertwined.

Acceptable and Unacceptable Use of GenAI in History Assignments

The list* below includes some specific historical tasks, with advice as to what is acceptable and unacceptable use of GenAI in your assignments, and when you need to acknowledge your use of GenAI tools.

The list is ordered roughly in the the order in which you might approach an assignment, not in the order of acceptable vs unacceptable use of GenAI. Make sure you read each section carefully before using a GenAI tool for any specific task.

For the reasons noted above, even when GenAI use is acceptable, it may not be desirable. It is often better to learn to do something yourself before delegating it to a machine, because this enables you to better judge what GenAI is telling you. Therefore the list also includes recommended alternatives to the use of GenAI.

If in doubt, don't use GenAI for a task without checking with your seminar tutor first.

*This list is adapted from the American Historical Association, ‘Guiding Principles for Artificial Intelligence in History Education (5 August 2025)

Keeping Records of GenAI Use

It is good practice to keep a record of any dialogues you have with GenAI chatbots in the course of your academic work. This is an extension of ordinary scholarly note-taking. Like other note-taking, it helps to organise your ideas, remember what you have learnt, ensure that your quotes are accurate, and trace any errors to their source.

It can also help to demonstrate intellectual ownership of your work, if you are suspected of Academic Misconduct. For example, it can help to show that you have used GenAI in an acceptable rather than an unacceptable way. You are advised to keep a record of any conversations with GenAI tools about your academic work, either by saving screengrabs or by downloading the transcripts of the chat history.

GenAI and Academic Integrity

All students should be able to demonstrate intellectual ownership of their work, and are responsible for the work that they submit for assessment. This is how we maintain academic integrity. A failure to follow these principles that gives you, or has the potential to give you, an unfair advantage is called “Academic Misconduct.” You are strongly encouraged to read the History department’s policy on Academic Integrity. Using GenAI in ways that this handbook has noted are unacceptable, is Academic Misconduct. You can find specific examples in the list of Acceptable and Unacceptable Use of GenAI in History Assignments.

A failure to disclose the use of GenAI, or the use of a misleading description of its use, may have significant consequences for your studies and may be prejudicial in any later Academic Misconduct investigations should they arise. As a result, you are advised to keep good records of GenAI use, as noted above.

Let us know you agree to cookies