AI tools for education

Choosing AI tools

Warwick AI tools

University provided digital tools have enterprise agreements to protect the data used and the security of staff. Warwick data should only be used in systems and tools provided to staff. Copilot is the recommended tool and no special licence is required.

Go to Microsoft Copilot

Types of AI tools for teaching

Distinguishing between categories of tools can help educators to consider in which contexts they might choose to use particular kinds of AI tools for teaching.

Narrow AI

All current AI systems are narrow AI - specialized systems designed for specific purposes. While some like chess engines focus on a single task, others like ChatGPT can handle multiple types of tasks, but they're all still specialised systems operating within defined boundaries not general-purpose intelligent systems (AGI).

Stand alone vs integrated

AI systems can be accessed in two main ways: as standalone services or integrated into existing software. Standalone AI services (like ChatGPT or Claude) require users to visit specific websites or apps. Integrated AI works within familiar software - for example, Microsoft's Copilot provides AI assistance directly in Word, Excel, and PowerPoint. This integration means students and teachers may be using AI tools even when they're not deliberately seeking them out.

Open-source vs closed source

AI systems can be either open source or closed source. Open source AI can be freely accessed, examined, and modified by anyone, making it potentially more affordable for institutions but requiring technical expertise to use effectively. Closed source AI is owned by companies and can only be used according to their terms - for example, ChatGPT and Claude are closed source systems that are easier to use but may have costs or usage restrictions that affect classroom adoption.

Approaches to selecting AI tools

There are many kinds of tools available but ultimately your choice of tools should be driven by pedagogical considerations, i.e. what do you need students to be able to learn or do? This should be at the heart of your decision-making even if you are selecting a tool for your own use to create learning/teaching materials.

Start with deciding on the use case, for example:

I want students to get assistance understanding readings/videos prior to interactive class meetings.
I want students to get assistance searching for source materials for a research project.
I want students to engage in a simulation - so I'm need a tool which can generate short talking head videos from photographs using a script.

Once you are clear on your use case, search for relevant tools to support the activity.

Like choices of word processor, web browser or search engine, sometimes there are families of tools that have different interfaces and functions. You may want to “test drive” a couple of tools in a family (eg. video summary tools, essay explanation tools, tutoring tools, etc.) or seek out reviews and comparisons of the tools to assist with determining which best suits your context and your intended use case.

Finding AI tools

The following are some suggestions to assist with looking for possible tools to incorporate into your teaching:

The following aggregator websites provide overviews of a range of AI tools:
A simple Google or Copilot search for “AI tools for education” will provide a further range of possible tools to consider.

Things to consider when selecting AI tools

If you are straying beyond Copilot there are some important things to consider to ensure that AI use is responsible, ethical, and sustainable.

Cost and sustainability

Are there free or trial versions that you can test?
Will the free/trial version work at the scale of your learning/teaching application? e.g. can students produce the work required using the number of 'free' uses.
Do you have a plan should the free version move behind a paywall?
Is the investment of effort (both learning and then teaching students how to use the tool) worth the return? Does this hold if you need to discontinue use due to changes in licence?

Data privacy and ethics

Remember no internal Warwick data, personal data or copyrighted materials should be given to an AI to train on and become publicly available.

Have you read the conditions of use and do you understand what happens with the data generated and how the data of others is used through this tool?
Will students need to share personal data to access, e.g. by setting up an account? What will you do if they do not want to do this?
Are you aware of the potential biases and data privacy involved in using this tool?
Are there any ethical considerations that you need to inform students/staff about?
Are you aware of potential risks when directing students to public AI platforms? For example, some platforms allow users to create and share AI characters/chatbots that may expose students to inappropriate or harmful content.
Does the tool have features that help support academic integrity? (For example, does it keep logs of student interactions that could help verify appropriate use?)

Accessibility and technical requirements

Is the tool compatible with assistive technologies?
Are there alternative options for students who can't use the tool?
Does the tool require specific browsers, devices, or internet speeds? For instance, some AI image generation tools require significant processing power and may not work well on older devices or slower connections.
Will it work reliably on university networks/systems?
Will the tool work with the types of content formats you need to input and output? For example, can you save responses as PDFs, export in accessible formats, download generated images in different file types.

Support and integration Quality and evaluation Geographic considerations

Making responsible choices

ALT's Framework for Ethical Learning Technology

This website presents the Association for Learning Technology's Framework for Ethical Learning Technology (FELT) designed to support the ethical use of learning technology.

Ethical approaches to EdTech

This website from the City University of New York provides guidance in thinking through the ethics of selecting educational technologies.

Striking a balance: navigating the ethical dilemmas of AI in HE

This Educause article explores the wider context of AI adoption, but includes some useful questions that we can ask when considering AI tools.

Using AI tools

The chatbot interface for many AI systems is based on a dialogue interaction. The ways you ask and the context included in the request shape the quality of the responses. If the response is not meeting your needs then expand, clarify or reformulate the request.

What is a prompt?

The way in which you ask an AI to do something for you is called a 'prompt'. Watch this simple explainer from Copilot: 'What is a prompt?' video [2 mins]. Copilot also has useful information on how to create a prompt and how to get better results with Copilot.

Prompt types

There are different ways to prompt, they yield different outcomes and reflect different levels of engagement when working with an AI. As Generative AIs become are refined the need for precision prompts has reduced. AIs are designed to understand natural language, so you can talk to them normally. Starting with a simple, clear request and adding more details or examples as you go along often works well. If you are building an API the quality of your prompts will matter - over-engineering prompts for casual use often just adds unnecessary complexity.

For complex analytical tasks having the AI walk through their reasoning step-by-step (chain of thought prompting) can help you to verify the logic and catch any mistakes, and can be particularly useful in specific cases, e.g. debugging code, solving mathematical problems.

Table of approaches to prompt writing - when to use, examples, and benefits
Approach	When to use	Example	Benefits
Simple direct request	basic tasks like summaries or translations clear, single-purpose requests when you want quick answers	"summarise this journal article about machine learning"	fast and concise responses less room for confusion natural conversation flow
Example-based (Give patterns to follow not to avoid. Use consistent formatting across examples. Experiment with the number of examples).	data formatting tasks specific writing styles custom output formats	"Format these references in Harvard style. Example: "Smith, J. (2024) Title of Book. Place of Publication: Publisher'"	ensures precise output format reduces back-and-forth clear expectations
Step-by-step	complex problem solving mathematical calculations debugging code analysis tasks	"Let's analyse this dataset step by step. Show your statistical reasoning at each stage" "Let's analyse how sustainable transport initiatives could work in city centres. First, outline the current challenges. Then, examine successful case studies. Next, identify key stakeholders and their needs. After that, explore potential barriers to implementation. Finally, synthesize this into practical recommendations. Share your reasoning at each stage." "Let's debug why this data visualization isn't displaying correctly. First, examine the data structure we're receiving. Then, check how we're processing this data in our code. Next, review the visualization library requirements. After that, test our data transformations at each step. Finally, verify the visual parameters we're setting. Show your thinking and any console logs at each stage."	verifiable logic easier to spot errors educational value better for complex tasks
Template based	recurring document types standardised reports consistent formatting needs generalisable role	"Write lab notes following this structure: [Methods], [Results], [Analysis], [Conclusions] "Act as a Socratic tutor following this pattern: [Ask probing questions about student's current understanding], [Let student explain their thinking], [Guide with targeted follow-up questions], [Help student identify their own knowledge gaps], [Encourage student to propose solutions]. Only move to the next step when the student has fully engaged with the current one."	maintains consistency saves time on repetitive tasks ensures all required info is included
Interactive refinement	creative writing complex analysis iterative development	"Review my draft abstract and suggest improvements. [After review] Help me identify areas where I could be more concise and better highlight my methodology." "Help me understand how social media affects political polarization. [After initial discussion] Let's explore what role echo chambers play in this. [As understanding develops] Can we examine some specific examples of how algorithmic content selection influences this process?"	allows for better iteration better final results more control over output can support process of learning as well as product.

Prompt strategies

There are many ways for formulating prompts, including ways to set the context and guide the outcomes of requests.

The basic components of a prompt include specifying what you want/want to do (required), providing contextual information (optional), giving system or style instructions (optional) and offering examples (optional).

Google’s PARTS heuristic Link opens in a new window(PDF) offers one way of conceptualising the key parts of a prompt:

Persona: Identify your role
Aim: State your objective
Recipients: Specify the audience
Theme: Describe the style, tone, and any related parameters
Structure: Note the desired format of the output

For further information here is a short video simple guide to effective prompt writing [Link opens in a new window9 mins] produced by an academic through the Thinking in public outlet.

Adobe Link opens in a new windowoffers related advice for crafting image prompts.

Prompt libraries

More useful things: prompt library

Created by Ethan and Lilach Mollick this prompt library includes:

instructor aids to help educators with preparation and teaching;
student exercises;
prompts for other uses beyond the classroom

Anthropic prompt library

A library of prompts for you to try if you are building an API via the Claude Console.

Copilot prompt gallery

A library of prompts to help you get started, with lots of examples to try or change to suit your needs. Some editable parts are obvious, denoted by a pair of square brackets, like [topic], [file], and [your title]. But you can also edit various parts of those prompts, such as the goal, context, expectations, and source, to suit your purpose.

Acknowledging use of tools

Using AI responsibly includes always openly acknowledging and explaining where and how AI has been used. Educators should model desirable behaviours by being transparent about their AI use.

Using AI in learning and teaching < Previous Next > AI and assessment