University of Edinburgh OER Policy update – Copyright, AI and OER

The Plaza
, , , ,
1

The relationship between Gen AI and the Commons is thorny to say the least, with copyright legislation in many jurisdictions struggling to keep up with the rapid pace of change and to address fundamental questions about the nature of creativity. What is it that distinguishes human creativity? What happens when Gen AI is used to help create a resource? Can it still be shared under open licence as an OER?

In order to provide advice and guidance for our staff and students, the University of Edinburgh has recently reviewed its OER Policy and included these recommendations about copyright, AI and OER:

UK legislation relating to copyright of AI-generated and AI-assisted works is ambiguous and evolving, and so for the present we recommend the following.

a. We encourage the use of CC0 for works that are AI-generated and do not involve a significant degree of human creativity. This means that no copyright applies.

b. AI-assisted works, which express the intellectual creation and creative choices of their authors, may be shared under Creative Commons licence as an OER.

For further information about copyright and Creative Commons licences, visit the Open.Ed website.

c. We recommend that AI tools are attributed whenever they are used. The University Library provides advice and guidance on Using Generative AI Tools in Academic Work, including citing and acknowledging the use of AI.

There remains a lot of concern about how Gen AI models are trained, and whether their outputs might infringe other creators’ intellectual property rights if they reproduce copyright works too closely. In addition, it’s not always easy to distinguish between AI-generated and AI-assisted works. How much human creative input is required before an AI created resource can be copyright?

To help address some of these issues, the University’s OER Service is also working on new advice and guidance about copyright, AI and OER to accompany the updated OER Policy, which we hope to share shortly.

Like all our open learning and teaching policies, the new OER Policy has been shared under CC BY-NC-SA licence and added to our award-winning collection of Open Policies for Learning and Teaching.

3 Likes
2

Hi Lorna, I’m excited to see the new University of Edinburgh policy on AI, it adds already to that tremendous collection.

Calling the situation “thorny” puts it mildly. Creating policies looks like roping jello (likely a misplaced metaphor), but as you are saying, it seems to undermine so many foundation layers of openness.

I’d like to think someone out there is collecting/ curating AI policies at the institution level, but my first searches were thin. There was an effort here in Canada but the links go to a page not displaying its intended WordPress content

I recommend often Lance Eatons collection of course level policies

He did have another collection for institutions but it has not been updated since 2024

I hope someone who might be reading this has better links, but I would love to some effort on a global collection.

Cheers Lorna!

1 Like
3

I asked Lance if there was a sizable collection of campus AI policies. He shared this from Campus AI Exchange which can be searched or explored by map

A few other interesting campus AI resources don the site

1 Like
4

Hello Lorna,

Thank you so much for sharing this. It’s really useful to read and it’s something I’m going to consider carefully. I wondered whether you would also share your news on the UK and Ireland OER Community of Practice mailing list? I’m sure you’ll get a lot of interest there too.

Kind regards,

Helen

5

And found Yet Another Spreadsheet of campus policy, well organized (thanks @NateAngell ) and quite a lot of other resources

1 Like
6

I was trying to be polite :wink: And yes, roping jello seems to be an entirely appropriate metaphor for developing policy in this area. It’s very slippery.

We do see this policy as being the starting point for more detailed guidance about copyright, AI and the Commons that the University of Edinburgh OER Service is going to provide. We actually have this guidance already written but it’ll be a while until it’s available as were migrating to a new website over the summer. I’ll share it here as soon as it’s available.

In the meantime the university already provides more general guidance about using AI in academic work which might be of interest:

1 Like
7

I’d be very happy to do that Helen. I’ll join your list and share it there shortly.

8

Love to see this! At OER Commons I’ve been thinking about how to address materials written by or using AI for a while now, and so far we’ve been relying on our existing collection policy’s criteria around quality to determine what gets added or not. We also tag things “AI Co-Authored” when we catch it, but I have big dreams to make it easier/more obvious for users 1) to self-declare in the resources they submit, and 2) to identify when/how resources are created with AI.

With thanks to @cogdog for having shared a bunch of ways folks are disclosing AI in their own works, I put together this resource that covers the options I’ve seen so far. For #2 above, I’d like to pick one of those and add it to resource overview pages, and then also work it into the ordinary user submission workflow. We just have to find the time/engineering capacity to make that happen. :sweat_smile:

3 Likes
9

Thanks Peter for sharing your collection of approaches and tools for disclosure of GenAI use. I can only imagine it might get more challenging to make a call to identify, can you say more from a librarian perspective how you do this on reviewing new submissions?

I’d like to think this can be the responsibility of those sharing works so you do not have to play the role of AI Detective.

10

Thanks Peter. The issue of being able to clearly identify whether and how AI has been used to create and contribute to a open resource is definitely not straightforward. This is some of the advice that Edinburgh University currently provides about using and attributing the use of AI in academic work:

The University provides guidance for staff on Using Generative AI in your Works which advises:

Be transparent about if/how you used AI. Never claim AI output as your own work.

This guidance also adds that it is unacceptable to

  • Present AI generated content as your own original work

  • Publish AI assisted content without appropriate attribution.

The University Library provides guidance for students, which is also useful for staff, on Using Generative AI Tools in Academic Work. This includes advice on citing and acknowledging the use of AI:

You should acknowledge all use of generative AI tools to help with any aspect of planning, writing or creating your assignment. Your acknowledgement should include:

  • Name and version (if included) of the GenAI system used; e.g. ELM; ChatGPT-3.5

  • Publisher (the company that made the GenAI system); e.g. University of Edinburgh, Edina; OpenA

  • URL of the GenAI system (for example ELM and ChatGPT)

  • Brief description (single sentence) of context in which the tool was used.

(ELM is the university’s own AI platform.). The IBM Research AI Attribution Toolkit that Alan shared here recently also looks promising.

1 Like
11

Very solid resources and approaches, Lorna. Your lists of items to include for acknowledgment reminded me a bit of the Creative Commons Guidance for Attribution via TASL (Title, Author, Source, License)… the University of Edinburgh approach might be NPUD (Name, Publisher, URL, Description of AI tool)??

This comes back to what I have seen more of in the past year in one a collaborative article I was part of and also just seeing in the posts/content I see educators creating where people use GenAI images and the only attribution given is “GenAI image” or “created in ChatGPT”

Let’s make NPUD a thing!

12

And also, because I am still coming across various collections if institutional AI policies, I saw this HEPI report by Sam Illingworth analyzing AI policies from 96 UK universities:

More than another list (although you can find it in the repo), Illingworth has shared in GitHub his methodology for running his analysis.

In there he mentions

Run this study in your country: The method is transferable. Everything you need is in this repository.

And it is all licensed CC BY

13

One more of interest is University of Leeds Traffic Light label approaches for course assessments:

The three categories of red, amber and green are not rigidly defined. They are intended to create a shared understanding between staff and students of how to use Generative AI tools in a particular assessment, by how much and at what stage of the assessment process.

  • RED: AI tools cannot be used
  • AMBER: AI tools can be used in an assistive role
  • GREEN: AI has an integral role and should be used as part of the assessment

Each color has an expanded definition, examples, and an icon to represent the level.