Skip to Main Content

Learning Technology Team

NILE Guides

Information for Staff

University position on Artificial Intelligence (AI)

UON recognises that Artificial Intelligence (AI) tools are increasingly embedded within commonly used applications such as Microsoft Office and web browsers. There are also a wide range of free and paid-for external AI tools available. In response to this, UON commits to supporting staff and students to become AI literate. Where appropriate, teaching and assessment may reference or use AI tools. You can find out more about the University's official position on AI using the link below.

About AI Design Assistant

The AI Design Assistant is a feature of Ultra courses, and may be used by academic staff to generate ideas and activities in the following areas:

  • Course structure;
  • Images;
  • Tests/quizzes;
  • Authentic assignment, discussion and journal prompts;
  • Rubrics;
  • AI conversations.

The AI Design Assistant only generates suggestions when it is asked to by members of academic staff, and cannot automatically add anything to NILE courses. Academic staff are always in control of the AI Design Assistant, and can quickly and easily reject any AI generated ideas before they are added to a NILE course. Anything generated by the AI Design Assistant can only be added to a NILE course once it has been approved by academic staff teaching on the module, and all AI generated suggestions can be edited by staff before being made available to students.

Please note that the AI Design Assistant cannot create teaching and learning materials (i.e., course content). While it can prove very helpful in the creation of activities to support teaching and learning (e.g., generating test questions from specific course content items to use as knowledge checks, formative tests/quizzes, self-test 'how am I doing?' type questions for students), it will not generate teaching and learning materials. It is also important to recognise that due to the inherent limitations of AI, anything generated by AI will need to be carefully checked by a subject matter expert before being approved for use in a course. 

AI Design Assistant has been created by Anthology in partnership with Microsoft, and you can find out more about their commitment to responsible AI via the links below:  

Benefits and limitations of AI Design Assistant

When using AI Design Assistant it is extremely important to bear in mind that it is a widely known and inherent feature of AI that it cannot get things right all of the time. Therefore, any suggestions made by the AI Design Assistant must be checked for sense and accuracy by staff who are subject matter experts and who understand in detail the teaching, learning, and assessment strategy in use on the module.

What the AI Design Assistant can do:

  • Provide ideas about how you might like to structure your Ultra course by auto-generating learning modules;
  • Offer AI generated, royalty free images to use for the course banner, learning module thumbnails, etc., in your course;
  • Generate ideas for rubrics;
  • Propose test questions (e.g., multiple choice/multiple option questions) based on a specific document or part of your course, or tests based on the module name and description;
  • Propose ideas for assessments, discussions, and journals;
  • Engage students in AI Socratic questioning conversations on themes or about subjects set by staff.

What the AI Design Assistant cannot do:

  • Create teaching and learning content;
  • Be relied upon to create 100% fault-free, factually accurate, unbiased suggestions;
  • Be used safely by staff who are not subject matter experts and who do not know the module in detail;
  • Understand the details of the teaching, learning, and assessment strategy;
  • Appreciate the wider context in which your module operates, including its relationship to the overall programme(s) in which it exists.

The most important thing to bear in mind if and when using AI Design Assistant is that it has no understanding of what it is doing, and that it is merely generating ideas/suggestions that you may or may not like to use. No suggestions can be generated by AI in a NILE course without being asked for by a member of academic staff, and no AI generated suggestions can appear in an Ultra course unless they are first checked, refined if necessary, and approved by the member of staff who generated them.

Complexity levels

There are ten levels of complexity for AI-generated content:

Complexity Description (US) Approx. UK Equivalent
1 Early primary school KS1 (Primary School, Ages 5 - 7)
2 Late primary school KS2 (Primary School, Ages 7 - 11)
3 Early middle school KS3 (Secondary School, Ages 11 - 12)
4 Late middle school KS3 (Secondary School, Ages 13 - 14)
5 Early high school KS4 / Levels 1 & 2 / GCSE (Ages 15 -16)
6 Late high school Level 3 / A Level / Foundation (Ages 17 - 18)
7 Undergraduate lower division Level 4 - 5 (Undergraduate Years 1 - 2)
8 Undergraduate upper division Level 5 - 6 (Undergraduate Years 2 - 3)
9 Graduate level Level 7 (Master's)
10 Advanced PhD level Level 8 (PhD)

AI Design Assistant tools

Before using any of the AI Design Assistant tools, please ensure that you are familiar with the limitations of the tool (see 'Benefits and limitations of AI Design Assistant' above), in particular, the fact that all auto-generated content must be checked for accuracy and bias before being used in a NILE course.

This is auto-generated content and needs to be checked for accuracy and bias.

For more information about the limitations of AI, in particular the generation of inaccurate information and the reproduction and amplification of existing biases and stereotypes, see: When AI Gets It Wrong: Addressing AI Hallucinations and Bias

Course structure suggestions / Auto-generate modules

AI Design Assistant can offer suggestions about the structure of NILE courses by auto-generating learning modules. The auto-generate module feature will create a number of learning modules at the top level of your NILE course, adding titles, images, and descriptions the learning modules it generates. You can choose how many learning modules you want the design assistant to generate, the complexity (i.e., difficulty) of the course, the prefix (e.g., week, topic, unit, etc), and whether to auto-generate images and descriptions for the learning modules.

By default, the design assistant will only use the NILE course title when generating course structure suggestions, but if you add in more information about the course into the description field - e.g., information from the module specification such as the module description, indicative content, module learning outcomes - the auto-generated modules will be more relevant. The ‘Select course items’ option can also be used to specify exactly which resources should be used when auto-generating learning modules. If ‘Select course items’ is used, the auto-generated learning modules will be based only upon the items selected.

You can re-generate the learning modules multiple times, so do not have to go with the first set of suggestions. You can also edit the auto-generated modules, so if you just want a number of learning modules with images in your course, you might find that auto-generating them is a quick option, and after they are generated you can add in your own titles and descriptions.

To auto-generate learning modules:

  • Select the + (plus) button at the top-level of your NILE course, and choose 'Auto-generate Modules;
  • Update the information under 'Define Learning Modules' and generate a new selection of learning modules if you want a new selection of modules to be created;
  • Select which learning modules you want to import into your course, and choose 'Add to course';
  • Once imported into your NILE course, you can then edit the titles, descriptions, and images in the learning modules.

AI image generator & Unsplash image library

Wherever you can add an image in an Ultra course, as well as being able to upload an image from your computer, you can also add Unsplash images. And, in many places in Ultra courses, you will be able to add AI-generated images too. Unsplash is a library of photographs, whereas AI-generated images are unique images created from a description. Both Unsplash and AI-generated images are freely available to use in NILE Ultra courses.

To add an Unsplash or AI image to your Ultra course:

  • Choose the 'Add image' icon, and change 'Upload from Device' to 'Stock Images from Unsplash' or 'Generate images';
  • Search for or describe your image, refining your search/description as necessary.
  • Once you have found an image that you want to use, select it and choose 'Next'.
  • From the following screens you can choose the aspect ratio, zoom level, and can set the display name, alternative text (this will be automatically provided if you are using Unsplash images), and whether you want your students to be able to view and/or download the image.

When describing an AI-generated image, you a providing an instruction to an AI image generator to create you a unique set of images based on your description. Every time you generate the images, they will be different, even if you use precisely the same description. The more information that you provide in the description about what you want your image to look like, the more likely you are to get an image that is close to one that you want. As well as describing the content of the image, you can also add information about the style of the image too. The following screenshots show how varied AI-generated images can be. Each used the same content description, 'deep in thought', but in the second and third screenshots additional information about the style was added.

Test question & question bank generation

There are two ways to generate test questions using the AI Design Assistant:

  1. From directly within the test tool - this method is best if you want to create a test where the questions will only be used in the particular test that you are creating. This is the quickest way of creating a test, and while this method can be used for complex summative tests, it is also very good for creating quick knowledge checks and formative tests based on specific content items or folders of content items.  
  2. Via the question bank tool - this method is best if you want to create large numbers of test questions which you intend to reuse in various tests, and/or where you intend to use the question pool test type to select questions at random from a large pool of questions so that each person taking the test is given a different set of test questions.

1. Generating test questions from within the test tool

Ideas for Ultra test questions can generated from within an Ultra test. The auto-generated test questions will use the name of your NILE course as a prompt, but adding additional information to the description field and re-generating the questions will make them more specific. The ‘Select course items’ option can also be used to specify exactly which resources should be used when auto-generating test questions. If ‘Select course items’ is used, the auto-generated test questions will be based only upon the items selected. The 'Select course items' option may be especially useful for staff wanting to create multiple tests, each based on one or more specific content items in the course, or for staff wanting to create a longer test built up from multiple auto-generated questions from different sections of the course, thus ensuring that the test represents questions testing students' knowledge from across the entire course.

Once generated and added to the test, the auto-generated questions can then be manually edited. Auto-generate test questions can be used multiple times within the same test in order to build up a longer test composed of multiple questions of different types from various different prompts and sources.

To auto-generate test questions:

  • Create a new Ultra test in your NILE course, or go into a test that you have already created and want to add more questions to;
  • Add a new question to your test, and from the question types menu, select 'Auto-generate question';
  • In the 'Define questions' fields, add more information into the description field to make the auto-generated questions more specific. Use the 'Select course items' option if you want to define which content items you want the auto-generated questions to be based on. Choose your question type (or leave as 'Inspire me' to get a random selection of question types), and number and complexity of questions. Then, re-generate your test questions;
  • Choose the questions that you want to import into your test, and select 'Add to Assessment';
  • You can now manually edit and save the questions that they are exactly as you want them.


2. Generating test questions via the question bank tool

Ideas for banks of test questions can generated from within the Ultra question bank. The auto-generated questions will use the name of your NILE course as a prompt, but adding additional information to the description field and re-generating the questions will make them more specific. The ‘Select course items’ option can also be used to specify exactly which course resources should be used when auto-generating test questions. If ‘Select course items’ is used, the auto-generated test questions will be based only upon the items selected. The 'Select course items' option may be especially useful for staff wanting to create multiple question banks, each based on one or more specific content items in the course, or for staff wanting to create multiple banks of questions each built up from multiple auto-generated questions from different sections of the course, thus creating a large collection of question banks from across the entire course.

Once generated and added to the question bank, the auto-generated questions can then be manually edited, and auto-generate can be used multiple times order to build up a large collection of question banks which can then be used in one or more Ultra tests using the question pool question type. The question pool allows a specified number of test questions to be drawn from a larger pool, meaning that each time the test is taken it will be composed of different questions.

To auto-generate banks of test questions :

  • From the course content area of your Ultra courses, under the Details & Actions menu, select 'Question Banks: Manage banks';
  • From the Question Banks page, select the + (plus) button and choose 'Auto-generate';
  • In the 'Define questions' fields, add more information into the description field to make the auto-generated questions more specific. Use the 'Select course items' option if you want to define which content items you want the auto-generated questions to be based on. Choose your question type (or leave as 'Inspire me' to get a random selection of question types), and number and complexity of questions. Then, re-generate your test questions;
  • Choose the questions that you want to import into your test, and select 'Add to Assessment';
  • You can now manually edit and save the questions that they are exactly as you want them.

Authentic assignment, discussion and journal prompts

Ideas for assignments, discussions, and journals can be auto-generated when setting these up in an Ultra course. The assignment, discussion, and journal prompts are designed to encourage students to engage in higher-order thinking, to apply their knowledge, to justify or challenge their beliefs and ideas, and emulate real-world situations.

To auto-generate ideas for an Ultra assignment, discussion or journal:

  • Create a new Ultra assignment, discussion or journal;
  • If you know roughly what you want your assignment/discussion/journal to be about, give it a title, as this will refine the auto-generated prompts to the subject indicated in your assignment/discussion/journal title. If you want to see the widest range of prompts, leave the assignment/discussion/journal title blank. With the assignment/discussion/journal title blank, the context for the prompts will come from the course title.
  • If you want your auto-generated assignment/discussion/journal prompts to be based on specific parts of the course, use the 'Select course items' tool to limit the auto-generation to these items only. This will ensure that the auto-generated prompts are only based upon the selected items.
  • Select 'Auto-generate assignment/discussion/journal. Choose your cognitive level (based on Bloom, et al's, taxonomy) and complexity, and whether or not you want the assignment/discussion/journal title to be auto-generated too. Add any further information to the description, if you want to further refine the prompts, and re-generate the prompts.
  • Once you have generated an assignment/discussion/journal prompt that you are reasonably happy with, choose it and select 'Add' to add the auto-generated text to your assignment/discussion/journal prompt. Once you have added it to the assignment/discussion/journal you can edit it so that it is exactly how you want it to be. 

Rubric generation

Ideas for Ultra rubrics can generated from the gradebook settings panel. By default, the AI generated rubrics will be fairly generic, but adding detail about your assessment to the description field and re-generating the rubric will make it more specific to your assignment. Once generated, the AI-generated rubric can then be manually edited.

To auto-generate a rubric:

  • Access the gradebook in your NILE course and select the gradebook settings cog;
  • Scroll down to the bottom of the gradebook settings panel, and under 'Course Rubrics', select 'Generate';
  • Choose which type of rubric you want (percentage, points, etc.), how many rows and columns you want, how complex you want it to be, and add more information into the description field. Then generate your rubric again to make it more specific to your assignment;
  • Once generated, you can manually edit the rubric, change the rubric title, and update the points or percentages, so that it is exactly as you want it.
Tip
If you add the headings and learning outcomes that you are assessing into the description field and set the number of rows and columns to match the number of learning outcomes, the auto-generated rubric will be more closely aligned to your leaning outcomes. For example, if you added the following to the description field, the auto-generated rubric would use the supplied headings and learning outcomes when creating the rubric:
Headings: Distinction, Merit, Commended, Pass, Fail.
Learning Outcomes:
[A] Clearly state the conclusion of an argument;
[B] Support a conclusion with a simple set of reasons;
[C] Develop a complex argument by using multiple reasons and intermediate conclusions.

AI conversation

An AI conversation is a Socratic questioning exercise guided by AI which lets your students explore their thoughts on a topic.

There are two elements to an AI conversation activity:

  • AI conversation: This lets students think critically about the topic chosen by academic staff;
  • Reflection question: This asks the student to share their thoughts on the activity. The reflection question also lets students flag any bias or errors from the AI as part of Anthology's Trustworthy AI Approach. Reflection helps students to better understand the responsible use of AI.

To create an AI conversation:

  • Go to the create Item dropdown by selecting the + (plus) sign anywhere on the course content page;
  • Scroll through the item type options until you reach 'AI Conversation' under 'Participation and Engagement';
  • Enter a topic of conversation in the content field, ensuring that your topic is open-ended and doesn't have a right or wrong answer;
  • Select a persona for the AI conversation. You can provide your own image, use Unsplash, or generate an image for the persona's avatar.
  • Enter a name for the persona and briefly describe the persona occupation, personality, or mood. You can also adjust the complexity of the persona's responses.
  • Select 'Save' when you're finished. You can preview the AI conversation as a student and interact with the AI persona by selecting 'Preview chat'. Previewing allows you to check that the AI persona is behaving as expected. You can edit the reflection question to guide your students on how best to reflect on the activity.

On submission, you can review the AI conversation transcript and your students' reflections. Please note that the AI conversation is a formative assessment by default, but you change this if you want to.

If you require assistance with NILE, please contact LearnTech Support using the NILE Enquiry Form

If you require assistance with PebblePad, please contact PebblePad Support using the PebblePad Contact Form