 So the one, two, three of data sharing agreements. This webinar is being brought to you by the Australian Research Data Commons, the ARDC. For those of you who don't know us, we have a purpose to provide Australian researchers with a competitive advantage through data with a mission to accelerate research and innovation by driving excellence in the creation analysis and retention of high quality data assets. Before I begin, I would like to do an acknowledgement of country. So we acknowledge and celebrate the first Australians on whose traditional lands we meet, and we pay our respects to their elders past, present and emerging. And for me, that is the Gadigal people of the Eora nation here in Sydney. So firstly, some welcomes and introductions. So welcome to all of you. Thank you for coming along to this webinar. It's great to have so many people on board and to see the interest that is out there within data sharing agreements. My name is Robin Burgess and I'm the research data specialist in data governance at the ARDC. And I have a real interest in the ways in which we're able to share and reuse data. And I've had many questions since joining the ARDC around the creation of data sharing agreements, which brought about the development of the documents that I wish to share with you today. And looking at the potential kind of one, two, three key areas that I think are important when it comes to data sharing agreements. Also, this is an opportunity to hear from those who have created data sharing agreements and have been working with them. So we've got Paula and Jake from the UNS from UNSW and Thomas from Swinburne University. He'll be giving their examples. And any of you on the call, if you'd like to share your experiences in chat or any comments you have that would be great. Also, hopefully there'll be time for questions. We'll do questions at the end of all the presentations. And I'm keen to also look at determining what next steps in the area could be and any further support you might be looking for from the ARDC. So do use the chat function as a way to for conversations. First, what is a data sharing agreement? I'm sure you're pretty much all across what these are having potentially worked with them. So they can be seen as a legal document and an agreement between individuals or groups wanting to share or access data. There are tool and document that will help protect the misuse of data from a project. And it can also be said that they're very much tailored for a specific project or organization. There really is no single template for a data sharing agreement. I've added an example here from the ONDC who have been working very heavily with data sharing agreements and they've produced a template to support with the access and sharing of government related data. So they will be running a session on data sharing agreements as well. These slides will be shared with you so you'll be able to click on the links that are within the present. So data sharing agreements guidelines is the first one that I want to present to you. So within this document. We look at commenting on what a data sharing agreement is, when data sharing agreements are needed, common types of data sharing agreements, the components that consist within the data sharing agreement and how to go about creating a data sharing agreement. So I just want to point out some of the key components of a data sharing agreement because this is often been a question that I've received. So we can very much look at ideas around the information about the data, the parties and the context of the agreement. So this is looking at details of the data to be shared, the purpose of sharing the data, any time periods affiliated to the sharing of the data and roles and responsibilities. The next section is related to the actual conditions of sharing. So this relates to what the data could potentially be useful and if there's any constraints in place with the use. Also sensitivity, security and retention of the data. Commentary on the methods of how the data will be accessed and actual details around the data such as formats. And then there can be some final sections and aspects related to the agreement to include such as a simple title for the document. So if you want to include aspects related to licensing or copyright conditions, you might have reporting requirements once the data has been shared, and it's also good to comment on risks and variations that might have occurred in the data prior to sharing it. So the second document that the ARDC has produced is related to data sharing policies. So within this document we comment on the types of data and the agent, the agencies have in their custody. The principles and strategies around sharing of that data commentary on governance and management procedures, when the policy might be needed, and once again key components of the policy. So comment on some of the key components that we list in this document. So the first is related to regulatory requirements, whether there's any other policies that are already in place that impact on your data sharing such as a research data management policy or open access policy. Also commentary on data governance require requirements are important. So this looks at the idea of ethics, roles and responsibilities and access to the data. It also touches upon the actual procedures for access in the data. So how will the data be accessed and what is the data request procedure. Other aspects include licensing and change management, and also a technical framework might be applied, looking at what technology and tools are out there that will support the data being shared. And also importantly with regards data is understanding the classification of the data, whether it's sensitive data or freely available data. So now we'll just a quick comment on the 123 of data sharing agreements. So these are the three key areas that I feel are very important when it comes to understanding and the application of data sharing agreements. So the one. So the first aspect that I'm, I would say you need to consider is do you actually need a data sharing agreement. So you might actually have data that you might actually just be able to make available through a simple license, or there might be another agreement process that could be in place for the sharing of your data. The second one is to review and understand the components of the data sharing agreement. So those are the areas I've commented on in this presentation so far. And to look at the creation of a template to support this and to subsequently complete the required sections. And finally one that I think is really important is that of getting support for the completion of the data sharing agreement document. This support might come from your legal office at your university or organization, and also you should also have many discussions with those for whom you were wanting to share the data with or receive the data from to ensure that there's compliance and there's compliance with regards the content of the data sharing agreement. And then finally ensure that the document gets signed off. And this is most likely to be done by legal entity. So just finally from me before we move on to the case studies. I'd like to ask you to kind of comment in the chat and I'll make note of any comments you make kind of what further support would you require with regard to data sharing agreements. Would you be looking for further guidelines, the idea of templates or examples. My final question is what sort of elements of data sharing agreements do you currently find the hardest, or is there an aspect that you find the easiest. So I'll be happy for you to share any thoughts on those questions within the chat. I'll stop sharing my screen, and I will hand over to Paula, Jake and Thomas, who will share their experiences with data sharing agreements. Thank you. And I apologize for the slight mishaps with the sharing and stuff that happened. So two of us from UNSW, myself, I work in the data governance area and Jake Simon who works in research technology and and research data management all the way. Jake explain when you get to him. Next slide please. So the data governance at UNSW, we've been doing data governance, first of all, high level, since about 2015 when take her others my manager, the chief data officer established the data governance framework. And then the work that came straight out of that was to put together the data governance policy suite. We've got quite an extensive data governance policy suite as a result. Policy sets out roles and responsibilities, including data ownership. We call it, we call our data owners, data controllers. So if I slip up and say data owner or data controller, that's what I'm talking about. And also other the role the policies also talk about quality and integrity classification and security. We've got a data classification standard. We have data handling guidelines which has just recently been superseded by a whole suite of cybersecurity standards. We've got data breach policy and procedure data retention procedure and a data sharing agreement procedure which is basically just formalizing what we've developed over the past few years. Then within that data governance policy suite we also have the data governance for research data. So Jake may touch on those and then related policies that we have privacy record keeping and cyber security. So all of those together guide what we do in terms of putting together data sharing agreements, which I'll touch on as we go to the next slide. So with data sharing at UNSW, we have a much more, I use the word mature, but it's probably got connotation. We've been doing data sharing agreements for teaching and admin data in a centralized way, much longer than we have for research data. But we're applying the things that work and are relevant in that area to research data as well as Jake will mention. But I thought it's worth taking you through what we've been doing with teaching and admin data. So, and we've got about 150 data sharing agreements in place for the teaching and admin data. So basically, whenever data is going to be used from one of our enterprise systems in another, or if it's going to be used for an enterprise system in one business unit by a different business unit, then there's a requirement for a data sharing agreement to be put in place. We've been doing this now since I think I've found some data sharing agreements that go back to 2017-2018. But I have to say that it's still coming as a terrible shock to some people within the organization when they hit one of our gatekeepers and are asked whether they have a data sharing agreement in place for this. So it's still a very organic process, but that's the basis on which we're putting data sharing agreements in place from one enterprise system to another from one enterprise system in one business unit to another. And then of course, if we, if we procure or use an external service platform system, what have you, then there must be a data sharing agreement between us and the provider of that service system platform. I mentioned gatekeepers. So we have policies. And as you all very well know, it's one thing to have policies and it's another thing to enforce them. And it's a learning experience for everyone. Over time we've established a set of informal gatekeepers that help us to enforce data sharing agreements. Of course, the data controllers, the owners custodians, the system owners, the data providers, cyber and the research ethics committee. All of those are that in a formal sense or an informal sense depending on the different policies and procedures that are in place. Say to someone who's requesting data, do you have a data sharing agreement for that? And if they don't, then they get sent to me. So I am the central point within university for dealing with data sharing agreements. They come to me or they use this, our SharePoint intranet to request the data sharing agreement. And then I get the process going. Next slide please, Jack. So it is a centralized process. I'm the hub and basically I then negotiate with and deal with all of the other elements that are spelled out here. So the form comes to me first and we ask them the basic information in high level terms. What's the data that you want? What's the source system? How are you going to transfer it? And how often what are your user and access controls? What are you going to be using it for? Are you going to distribute it any further? And how are you going to store it? Where are you going to store it? So sovereignty is obviously and storage sovereignty is obviously an issue. So the area that I work in on the data governance manager and myself and my boss are the data governance officer. Once I've gathered that first bit of information, then I will request additional information. I'll need a spreadsheet of every single data field that's required because I know that the data controllers will want to see that. We'll look at them all. I will help the data user classify all of the data fields. I will determine whether cyber has assessed either the solution that they're using in-house or the solution that's being procured. I will talk to the privacy officers to see if there's any privacy considerations either relating to the data itself or the procurement contract. And as I said, I'll talk to cyber and then I will draft together a data sharing agreement. We have a template, but I've discovered that the template only goes so far and that pretty much every situation requires a bespoke data sharing agreement. Having gathered all of that information, then we have an approval meeting. We'll hold it with the data user, the data controller, privacy officer, system owner. And essentially the data controller interrogates the data sharing agreement and the data user to get to their comfort that the data will be used appropriately for an appropriate purpose. We'll have all of the access controls in place. And the system, the data user understands that they can only use it for the purpose that's been approved in this agreement. Generally speaking, the data sharing agreement has been approved for three years. That's general. Sometimes only for six months, if for example, cyber is still doing penetration testing and is happy to give an interim approval, but not a long term approval. So, you know, the timing may change. Next slide. Thanks, Jake. Thanks over to you to talk about how we're applying this, these learnings into the research data sharing space. Yeah, thanks Paula. So I'm from research technology services. As Paula mentioned earlier, we have a function within that, which is the research data management single point of contact IDM at UNSW. I use just an email address really, but so that's been our main point of contact with people who want to know about how to do a research data sharing agreement. And because we've had some inquiries on that Paula actually wrote a sort of version of her teaching and admin data sharing agreement as a research one. So adding research things and changing bits and using some stuff from other templates that we found. So that's it on the right there. So we have that template. I don't know that I think we've had one person that asked for it and not sure if anyone's actually used it yet. So that was it's been around for a year or so now. I think our main interest main sort of contact with these data sharing agreements is that we don't see a lot of them we're assuming that there are there are some happening in the background but when we mainly know about them when they come to our email address when they have some sort of problem or they're asking questions about us and then we work with Paula usually to try and figure out what is happening and what we need to do to help them out. In my experience we've been doing this IDM at UNSW for years, something like that now and we haven't had too many data sharing agreement inquiries during that time. We have had some and we've usually what has been our experience is that people come to us with a data sharing agreement because they're asking for data from probably a government body health district, someone like that. And they've brought a template, their own data sharing agreement to UNSW that presented that to UNSW and somebody wants to know some details about how to fill it in. So usually it's not sharing agreement that we have there it's something that's coming from someone else. They have a pre filled agreement sometimes it's you know it's pretty good sometimes it's very vague or it's designed for companies to fill out or something like that. But we've helped people with fill them in there's some considerations there I think Robin covered a lot of this stuff earlier. But that's what we're trying to make sure is is detailed appropriately in the data sharing agreement. And often what happens is that the data sharing agreement is presented to us, we have we give it a read it's legal ease often. And so we have to fix the road and make sure that we're not promising things that we can't do. Because we're research technology services, I often are involved in the technical bits about data storage or systems that have to be used requirements that come from the external people about what systems were allowed to use, trying to match those with requirements that our own cybersecurity and other policies cover. So that make sure that we're not sort of conflicting but what is that that's ended up being is a negotiation, I find it is an interesting output of what that is is that they come with an agreement which looks like a legal thing. But it ends up being you know they say oh you have to use this system or that system and we can't comply with that because we're a university or we don't have access or the researcher doesn't have a budget or something and so we have to have a back and forth with the data controller or the data provider and figure out what we can do and what we're allowed to do. And most of the time that's been pretty good sometimes we have had a couple where it's gone it's gone away and never come back so possibly they haven't been able to get the data because they haven't been able to come to terms that we're agreeable with the data provider. The template that we've got here was created actually when a data provider hadn't done it before and was asking if we had a template that they could use to to form an agreement because they didn't really know how to do it and so we said okay this is what it looks like. So it was a thing that a template like with the sort of things you see there and then they helped them and the researcher that was asking for the data filled those out. We have had these are the sort of some of the places that we've seen data agreements happen. German Bureau of Statistics is common ATO we've had, there's a consumer data right page, which I've recently been involved with some people trying to get some data from there. So it's a very complicated and high requirements for security and systems and and certifications from that one which was interesting. So these people. So I think that's been the most of the things that we've got from sort of health districts. And now this consumer data right so they those people all sort of brought their own data sharing agreements that we were sort of asked to sign basically or the researcher was asked to sign. I just wanted to mention data place because Paul has been doing a lot of work with this and hopefully it's going to be somehow helping us with with data sharing agreements. It's just the Australian government website has not quite gotten online yet, but the idea being that you can sign your institution up for it and then the Institute government based data sharing stuff will go through there and we again we sort of have had a few people go through Paula. She can talk more about that if people are interested maybe but to try and use that a place for creating data sharing agreements or negotiating data sharing agreements. But at the moment, it's sort of half made so I don't think we have a lot to talk about there yet because it's not quite ready. So that's all I was going to talk about there. So maybe we can pass it on to the next person. Sure. That was fantastic thank you Paula and Jay for giving it as an insight into the areas that you're working on with regards data sharing agreements what's working what's not and the kind of logistics and the number of people who are kind of involved in the process. And you mentioned data place there yeah. The RDC have actually got a date place data accreditation meeting happening tomorrow at three o'clock if anyone's interested I can send you a link to that meeting it's a support group to discuss aspects around data place and gaining accreditation as a data user within the platform. And so at this point, as I mentioned we'll take questions at the end so I'll hand over to Thomas, and hopefully you'll be able to share your screen. Well let's have a go. All right so hopefully screen sharing shout out if I'm not looking good. I'm just more informal than the last presentation and I'm going to focus mostly on a little bit of the process and the outcome of a template that I developed. Ironically, before I had seen Robin's really nice data sharing agreement development documentation that's been really interesting, looking at the alignment between it and that document. I think part of the RDC institutional underpinnings project and I was running part of the policy development thread of that larger project within Swinburne. Part of that, initially, the aim was to just put together a policy that could be kind of modular and used between universities and adapt it and things like that around research data management. This came out of the fact that Swinburne did not have any research data management policy at all at the time which is, I think, a big oversight for a university with technology in the name. But as part of that project it became increasingly clear that actually there was a whole load of an ancillary material around it that was needed you know you can't just have a standalone policy and I think everyone knows that it. To have that you are inviting essentially zero percent compliance. So, there was also some guidelines and how tos and FAQs and training material developed around it, as well as a whole comms communication package around it, but also a few templates and that helped people out. The three sort of relevant templates from this point of view was a read me template to describe individual data sets or folders or directories of data set, a research data management plan template, and a data sharing agreement template. The reason behind these was because again we previously didn't have anything so we wanted to make sure that we were putting something out there that was relatively easy as a low bar to start with. So for example, our research data management plan that not in any fancy system or through a control form, they are Word documents, which many universities have previously had and then moved on from. This is, you know, starting from a relatively low bar, but the benefit, I guess, from the pointer from that point of view is that Word documents have quite a useful balance of being able to have, you know, controlled areas where people enter information, but also being highly adaptable. And then it's already been mentioned by basically everyone who's previously spoken. You can't have a strict data sharing agreement template, there's always going to be the particularities of the data set or the signatories that essentially mean that all of, even if you start out with a template it'll inevitably be significantly edited since then. So part of this I went out and I started sniffing around the internet to see if I could find some creative commons license sharing agreements because that would save me a whole load of trouble if I could just adapt an existing data sharing agreement template. I went into a reshare screen actually, because I ended up with this folder full of dark sharing agreements that I've called from around the internet. But interestingly, they kind of fell into two broad camps. One was data sharing agreements written in legalese so they look very much like legal documents with a large amount of text a large amount are typically organized into statements and section and usually kind of organized around big definitions throughout the text. Another category is to have grown out instead from the research data management plan sphere, just kind of the direction that I come from as well. And they look more more like research data management plans to be shared between more than one party, and then everyone essentially signs up for this joint research data management plan. Those are broadly the two camps I've seen things fit into and it's more or less 5050. But it's interesting that they're kind of coming from such different directions. Unfortunately for me none of them were created commons license all of the, all of the data sharing agreement templates that I could find out there online at least at the time this was back in 2022. So, you know, always reserved copyright so no dice and being able to talk up in that direction. So what instead I did, I'll just go back to my original share. Instead, I did was as part of this kind of package of material. I developed a new templates that I was then happy to make creative commons license so I just did the DIY link up in the chat so that people have it just in case it's useful to get access to that. So what I'm going to start up with is this document. So, I'll just quickly screen share it is going to be relatively familiar from anyone who's ever put together a research data management plan because that was the format I guess that we went for. Swinburne livery obviously because we're using it at Swinburne but it's adaptable. I'll just highlight the licensing information down at the bottom, which is that the template is licensed under creative commons by for license. So, if any university out there is in need of a data sharing agreement and wants to cut as a template and wants to use it as a starting point they're welcome to. I already linked to the DIY for the, but there's a no day repository. This is a live document. So, as I gain additional feedback from people I have been updating it over time it's now on version 0.2.5 I think. So it's going to be updated over time but it may be the very useful starting point for different groups. And it actually aligns reasonably closely to what's been talked about already in terms of the kind of logical sections for data sharing agreement to form. There's a few things that are maybe slightly different to it so I thought I would highlight those. The first thing is that this is meant to be flexible enough to be used for requesting data from a third party into to be used within the university and also licensing of data that originated within the university to be shared with a third party. Swinburne does do some medical research where we're interacting with hospitals and other clinical research institution, but we do a lot more research with private corporations and companies we've got a lot of engineering and technology work that goes on here so that tends to be the focus. And one of the things that comes up in that context is having to be quite specific about a project may have multiple data set. And one thing that can be slightly tricky is that not all of those data sets are going to require the same handling. And the other thing that we've come across is in clarity between the original raw data set being provided and derived data sets that are being generated from them. And so that was an all from those. And so that was another thing that we wanted to try and be quite specific around where possible. And so some of this is relatively obvious stuff, the dark date and date, the descriptions of those data sets that were all clear on which data sets we're talking about particularly the project can have dozens of different data sets involved. The usual things that side of things in terms of who has access to the data sets, but also information about, can these data sets be disseminated on to other parties. We have had problems, for example, where researchers have ended over a data set to an external collaborator and that external collaborator has used a platform or a tool, which automatically shares that data on the third party is all vice versa, which is particularly embarrassing from our point of view when we are exposing data to third party platforms unintentionally. And so it's just a work while being clear about whether that's intended or not. In terms of storage, actually, this is already a little bit out of date, of course, because our story is being commissioned by the end of this year so previously we were more or less the options were cloud store, Nectar, or other, particularly, particularly being asking researchers to be clear where it's being stored. Now that our stories being the commission we're likely moving over to using one drive for those same functionalities one and the sort of same functionalities. But in particular, being clear that from the university's point of view, that was being stored both encrypted and within Australian borders. Now this is obviously relevant from very domestic point of view, if we're talking about an international collaboration, it's entirely possible that a copy of that data would need to be stored externally outside of Australia. But again, from the point of view of university data, most Swinburne, for example, informed consent forms around human data will specify that the data is only going to be stored domestically within Australia because of different countries have very different laws around under what circumstances the government of that country can demand access to your data. If a researcher was generating a data sharing agreement for sharing data with a partner internationally, we'd want them to seriously think about what the solution is in that case, the two broad solutions being either okay in that case you're in the evening encryption data to be stored encrypted outside of Australia in which case you might need to check in with the ethics office to see whether that is sufficiently compatible with your informed consent. And with expectations of the legal systems of those other countries, or you can set up a system whereby the only copy of the data, it will still be stored within Australia, but external parties will be able to access that data, API into it or something like that without storing the local copy outside of Australia, for example. When we are talking about transfer of data, again, a common issue we've found within the research ecosystem is sharing data by email, right, and I think that that's pretty standard across most universities, the downside of that being of course that is not encrypted, which really puts a big hole in your security when you're moving something from one encrypted server to another encrypted server via an entirely unencrypted intermediaries sort of undermines the entire thing. Obviously, that doesn't apply if the files are being manually encrypted using some other system attached to an email, and then being received that's fine but that is not the typical way in which researchers interact with email attachments. So these were our standard encryption options within, within Swinburne, but again, the idea would be that it could be adapted to whatever the standard transfer methods are at different universities. And again, this has already been mentioned, you know, what's the, what's the intention around the destruction of these data that it is pretty common within a university research ecosystem or data to be retained way longer than it's actually intended to be. One of the issues around, you know, within the ethics office side of things is already often we're telling participants that data is going to be destroyed after a set period of time. We're actually enforcing that that data is destroyed 15 years after, after the project is not trivial. And the same goes in this case but at least it's important to flag the intention. And just the standard kind of periods of time that one might expect data to be retained for, including the concept of indefinite retention. Importantly indefinite retention doesn't necessarily mean permanent retention indefinite just means, no, I'm not going to tell you how long because I don't know how long but at some point I assume I'm going to delete it. And a statement that's going to be permanently retained is more in the lines of, you know, talking about open research data and things like that, or data that's going to be important in a patent application where you actually do have to make a commitment but no, no, no. We're going to keep this data actually safe and secure. And permanently archive it. And then the next section is much more in terms of the what is permitted to be done with that data set. So again, something that is common within the research sphere is researchers want to say, look, I'm happy to share my data with you, but only for certain purposes. I don't, this isn't carte blanche to do whatever you want with that data set. This necessarily has to be quite reform. There's so many, you know, potential users right you're not going to have a drop down menu or a tick box options for what's going to be done with this data, but there is at least suggestion as to what sort of things we mean in terms of what sort of derived data sets are you expecting to be using this data for what is the scope of the project what other data sets you're going to be combining this with. And similarly, again, we've talked about, well, what was the retention of the original raw data that we initially shared, but there's typically a large number of derived data sets that come from that data. So it's not really, you know, the spirit of a data sharing agreement right is not that you get to get a whole load of raw data, do some trivial manipulation to create a derived data set. In the organization that's okay I deleted the raw data and then keep your derived data set that sort of circumvent the entire process. So being clear around what we're expecting from derived data sets, because almost invariably, those are going to be retained by at least one party and often by both parties, but within a university ecosystem. It's not uncommon for us to, after they were actually the data that we get from you is going to be combined with certain facets of that data is going to be combined with other data sets we have in house, and derived data sets are going to be used as part of a publication, and as part of our responsibilities around that publication we're going to make that derived data set open access is very important to be clear about that because that can be a very different behavioral expectation compared to a lot of private industry where they. are indicated ahead of time. A lot of organizations outside of the university ecosystem may be very surprised by researcher intentions to be quite so public with they de identified or agglomerated or otherwise derived versions of a data set shared with them. So the next couple of sections of very researcher focused and again typically don't come up in terms of the enterprise data sharing agreements. And this is a little bit like, this is more kind of in common with research collaboration agreements that that some researchers do to kind of. prevent arguments over authorship later on. And so broadly it's talking about what what outputs are expected from this data and who's going to be authors on those research outputs, the sort of thing that it typically isn't relevant to. It's a matter and often isn't thought about from from external industry partners point of view but can be extremely front of mind from a researcher point of view. And then finally yeah we've talked about you talked about this offboarding deletion process, what's going to be done with these research outputs. Again, is the partner expecting oh yeah there will be a research output you are writing a report and you're going to give that report just to us. Because our expectation is very different to. Oh no actually we're going to be making a research output that actually going to be made public, either public and paywall or public and open access. So very occasionally some partnerships have expected that there'll be some sort of research output that's only going to be retained by one of those parties. I think that this is actually extremely rarely done but sometimes it's the expectation that a partner can have so again, like I've said a million times, it's just laying out expectations where possible ahead of time. So the human data section is kind of very specific to their kind of the ethics office side of data sharing agreements and then it's simply the last section is the signature section at the bottom, you know organization position name signature date. And very basic stuff so you can see how this is, you know, everything up until that is very much looking like a, an ethics application for a research project or a research data management plan within the university context, very little in terms of the legal of a data sharing agreement. It becomes a little bit more like that with the, with a reasonably long definitions section just so that everyone's on the same page as to what exactly we're talking about when we say sensitive or we say shared. The only other thing that is relevant in this context is I also put together a little kind of description around the logic of this data sharing agreement so if you're using this as the starting point within your own university organization institute whatever. You can see at least what the thought process was that within went into the way that it currently is. That's going to be useful, more useful for some settings than others but at least it gives an idea of what the thought process was behind it. I'm just going to have to cut in there. We're kind of nearing the end of time so if you've got some just final comments that you'd like to share and then can quickly open it up to see if anyone's got any burning questions for the panel. As you may have guessed from the fact that that's the end of the document that's also the end of the main things I wanted to say so I shall unshare now and open it up to questions in general. That was fantastic Thomas so good to see your template and all the steps that are involved and really good that you've added that commentary at the end of the document to support and how to kind of complete that and I think people will definitely have a look at that template to help support the development of their data sharing agreements. So I just want to open up the floor to the participants if anyone's got any burning questions for for the group. We've taken note of the commentary within chat around kind of areas of interest and areas of need. So any questions anyone would like to ask. I'll ask a question if no one else will. For any and all of the presenters. Have they been approached about AI training or model training for any of their data either the administrative data or, or otherwise. As an archive we have been requesting have many requests coming to us on those grounds and we're not sure that our regular contracts kind of cover that because they're not sharing the information again but they are utilizing the information that is tamed to train their model or train their AI. Kind of on sharing that information. Does that make any sense. So I was wondering if any of the three presenters have had similar and data sharing agreements. I'm happy to briefly comment so I have part of my background is in unsupervised machine learning so I haven't done that much supervised machine learning but broadly speaking. But the, you know when training a large language model for example or some other kind of supervised machine learning the current trend is to call it AI but it's kind of, it depends on the exact technique, or any other sort of neural network. I would say that the, the neural network that you end up with is a derived data set. So that's the sort of thing where you would want to be clear as to the methodology but the derived data set and how it's going to be used as the sort of thing that the original participants should be aware of because a lot of those sorts of machine learning models. You can't extract out of them the exact raw data but you can extract out significant aspects of that data. I can add that we've had a couple of data sharing agreements in the teaching and admin space that have been not exactly using AI but say scanning metadata or scanning IP addresses or that type of information. And certainly have been putting data sharing agreements in place because it's Thomas has said it's about who's using the data, what are they using it for and what's the output. So we have been putting those data sharing agreements in place for those types of scenarios. David you've got your hand up. Yes, thank you. Thanks Thomas that was a very detailed document. I saw there was a reference to authorship, whether authorship of a publication or document was part of sharing agreement. But if I remember well in the code for responsible conduct of research, only providing data is not enough to be an offer. Yeah, agreed. And so that's one of the things where I think that sort of expectation needs to be laid out ahead of time. Occasionally it is the case that as part of data sharing agreement there's also the expectation to share interpretation expertise it's not just the data, it's we're giving you the data so that you could see it and we will be collaborating on this. Okay, yeah. I think we're coming to time now so I would just like to thank the speakers for the time they've given to present today during this session. It was really informative really inspiring I would say I've learned a lot myself. I see a lot of various that I would like to kind of work on further to within the ARDC to support the universities and organizations with the concept of data sharing agreements. And it's been great to be able to kind of share the areas of interest and the work that people have been doing here. And this finally I just want to apologize about the technical difficulties at the start of this, sometimes zoom likes to have a mind of its own. So thank you once again everyone for coming along and look out for sessions in the future with regards this sort of information. Thank you again.