Online Learning as a Commons: Supporting students’ data protection preferences through a collaborative digital environment

3

Technology and education have long been integrated. From e-mails to using laptops in the classroom, technology has allowed for more flexible and inclusive ways of learning while introducing new methods for collaboration and information sharing. [2] However, technological developments have also increased the responsibilities that institutions have over student data, expanding and blurring the lines of what education data entails. Borgman describes education data as “grey data”, where teaching, learning, and administration activities have fallen within the remit of data collected by institutions. [3] As a result, Borgman argues that it has become more difficult to assess the risks and responsibilities associated with data collection, where the privacy frontier for institutions spans open access practices, uses and misuses of data, and curating data for privacy protection.

4

The digitalisation of education has resulted in greater data collection, storage, and analysis through learning analytics. While learning analytics can help institutions understand student engagement, improve teaching, and the overall student experience, [4] they have similar characteristics to big data and so have similar data protection concerns, particularly regarding relationships between universities and students. [5] As a result, in considering a data protection framework for governing learning analytics, Cormack argues that there should be two key stages for protecting student data. [6] Firstly, there should be greater ethical care on the discovery of significant patterns and must include safeguards for individuals’ interests and rights. Secondly, applying those patterns to meet the needs of individuals requires their informed consent or a contractual agreement. Prinsloo and Slade further create a framework to support learner agency, [7] recognising that it is impossible for individuals to comprehend the scope of data that might be collected, analysed, and used and its implications when it comes to learning analytics. [8] This framework includes contextual integrity of privacy and data, student agency and privacy self-management, rethinking consent, and employing nudges. Similarly, Sclater develops a learning analytics code of practice, with a methodology for setting up appropriate governance structures, developing a taxonomy of the issues, drafting the code, consulting stakeholders, and embedding it within institutions. [9] Models that incorporate privacy-by-design have also been considered essential to learning analytics systems development, where the learning analytics design space can address issues of privacy, identify means to control data, and support trust between education stakeholders. [10]

5

As education technology becomes more commonplace, HEIs have to identify and manage the challenges around the increase of data collection, analysis, and management. Given that the authors are based in the UK, we contextualise our assessment of the online learning landscape in the country. For example, organisations such as the Joint Information Systems Committee (“JISC”) and the Office for Students (“OfS”) have provided guidance and supported the creation of education digital infrastructure, services, and learning providers. These include reports on learning analytics, [11] lecture recordings, [12] and supporting students with disabilities. [13] HEIs also employ data protection officers and research archivists to meet regulatory requirements.

11

Institutions such as the Open University have long run courses with a strong online component [27]demonstrating how implementing online learning technologies can improve the educational experience with clear communication of how student data are used. [28] The rapid integration of new technologies for remote learning raises the possibility of data protection harms, introducing new concerns related to online learning and privacy. For example the data protection risks emerging from the use of online platforms, such as Zoom or Microsoft Teams, include the allocation of roles and responsibilities of stakeholders, transparency of data processing and possibility to effectively exercise data subjects’ rights, extra-EU data transfers, and the challenges of e-proctoring systems. [29] Universities’ adoption of cloud computing also has implications beyond individuals’ privacy, with questions of academic independence and integrity. [30] The data protection challenges that arise in the specific areas of lecture and tutorial recordings, e-proctoring, and platform ecosystems are discussed below.

15

In response to some of the digital and data-related challenges that have arisen from COVID-19, many organisations have looked at the impact of the pandemic on education. The OfS engaged stakeholders to produce guidance establishing the essential components of successful digital teaching and learning, recommending core practices HEIs can use to improve online learning for students. [49] JISC have written a report to understand the COVID-19 response and explore the future of digital learning and teaching. [50] Policy solutions were also devised for identifying the future role of emerging technologies in education and training. [51] The Open Data Institute has also suggested public engagement to support data governance considerations when working with online learning data. [52]

16

More broadly, the UK Information Commissioner’s Office established a code to help employers comply with the GDPR and to encourage them to adopt good practices, including monitoring at work. [53] The UK Department for Education created a COVID-19 addendum to acknowledge issues of privacy in schools. [54] The European Commission also identified the need to create a trusted digital education ecosystem with high-quality content, user-friendly tools, value-adding services and secure platforms that maintain privacy and uphold ethical standards as part of its new Digital Education Action Plan. [55]

21

Adapting the commons to individuals’ collective digital data, Hess and Ostrom developed the knowledge commons, where knowledge is the CPR. [70] As new technologies enable information capture, the knowledge commons recognises that information is no longer a free and open public good and now needs to be managed and protected for archival sustainability and accessibility. Crucially, the commons addresses data-related governance challenges that arise due to spillovers created by the reuse of data, thereby increasing its value over time. [71]

22

An example of a knowledge commons is a university research repository. [72] Developing a university repository requires multiple layers of collective action, coordination, and shared information and expertise. Academics and researchers can contribute to the repository as the more it is used, the more efficient it is to the university. Others outside that community can browse, read, and download the repository, further enhancing the quality of its resources. By breaking down large, complex, collective action problems into action spaces through the Institutional Analysis and Development (“IAD”) framework, [73] collective action problems can be assessed so that institutions can more accurately meet the needs of the community, including how information, knowledge, and data can be used to serve the common good. [74]

23

The commons has been further adapted to the university environment. Madison illustrates that as universities continue to evolve, the nature of the university may change from a knowledge to a data-oriented institution, resulting in the conflation of data as knowledge. [75] As a result, the way institutions may be governed could also change. In order for HEIs to manage their resources for maximum benefit and minimal social and private harm, HEIs could consider the knowledge commons to examine data governance beyond intellectual property rights and be open to multi- stakeholder engagement when creating university policies and meeting third-party obligations for education data. Although the risks of data collection, sharing, and security are not explored, Madison offers insights into how university data could be managed as a commons via strategies of openness, sharing, and polycentricity, but with contextually-appropriate elements of proprietary management and exclusivity with regards to intellectual property.

24

As a result, in order to increase student agency in protecting their personal data, a commons could be created to support collaborative means for them to meet their data protection preferences with the knowledge of their institutions’ data protection practices and of their individual data protection rights.

25

A data protection-focused data commons allows data subjects to collectively curate, inform, and protect each other and the collective exercise of data protection rights. A commons that focuses on data protection can provide students with agency over their personal data and redress the power imbalance between them and HEIs. For students, participating in the commons allows them to improve their understanding of their institution’s policy and external organisations’ guidance when it comes to collecting, processing, and sharing their online learning personal data. The commons also allows them to ask questions to experts, raise any questions about data protection to other students, review their consent decisions on tutorial recordings, and exercise their data protection rights. It simplifies the data protection rights procedures by including information, instructions, and templates on how rights should be collectively exercised, giving data subjects an opportunity to engage with and shape the data protection practices that govern how their personal data are protected.

26

Creating a data protection-focused data commons could help identify how much understanding and control data subjects have over their personal data, supporting them in choosing their data protection preferences. A commons for data protection does not require the creation of a new legal framework, but rather, operates within the current data infrastructures used by data subjects and acknowledges the limitations of existing laws, technologies, and policies that steward data. Thus, the focus on data protection as part of the data commons shifts data protection responsibilities away from the individual alone to their community, where knowledge, expertise, and experiences can be pooled together to identify working solutions. Although personal data are still kept personal and private, the collaborative nature of sharing, discussion, and advising on data protection problems opens up potential options for everyone to support informed decision-making and achieving data protection preferences through a data commons.

Figure 1: In a data protection-focused data commons (green), the data subject specifies to what extent they would like their data to be protected based on existing challenges pre-identified within the data commons for the use case (red). No prior knowledge of existing law, norms, or policies are required. Along with stakeholder information (blue), the data subject specification is then used to inform their data protection outcome as generated from the system. As the outcome is data subject-centred, decisions ensuring the protection of the data subject’s personal data may override existing preferences, policies, or standards set by other stakeholders. Data subjects can return to and review their outcomes, add their data subject experiences to the data commons, and participate in the co-creation process at any time.

27

In our previous work, [76] we interviewed commons experts to assess if and how a commons framework can be applied to data protection to support the protection of data subjects’ personal data. From those interviews, we found that collaboration across stakeholders and disciplines could overcome excluding data subjects and doubts about the effectiveness of the commons. The purpose of the commons needs to be clear because the use of the commons model is a choice, and that clarity allows for new iterations of the commons to best suit data subject needs. The commons must include the vision of communities and people about what is at stake, what it is about, how it works, and how data have been managed. Ultimately, commoning was identified as a verb, where the community has to actively participate in the development process and its application and is necessary for successful co-creation and participation. Based on these findings, we adapted an IAD framework and policy scaffolding for the creation of a data protection-focused data commons (included in Appendix A), which we now apply to create an appropriate commons for online learning.

37

To test the commons tool, we split the study into three parts: an entry questionnaire, an interactive task to test of the commons or control application, and an exit questionnaire. Figure 7 illustrates the different stages of the study.

Figure 7: Study walk-through summary

38

The study was conducted online between April 2021 and October 2021. As the authors are based in the UK, all participants were undergraduate and postgraduate taught students studying at UK-based universities over 18 years of age. The study was approved by the relevant University Ethics Committee. We outline the initial survey, application testing, and final survey components below.

42

34 students participated in our study. The participants studied Computer Science (6), Management (3), Finance (2), Philosophy (2), Psychology (2), and 19 other subjects that were only studied by one participant. 23 participants were undergraduates and 11 were postgraduates. Our participants predominantly identified as female (26) with seven males, and one not disclosed. From our results, we did not find any correlation between the discipline of study, level or year of study, or gender.

43

Regarding tutorial recordings, 19 participants thought that they had control over whether a tutorial was recorded, with 10 disagreeing and five were uncertain. When asked about the university’s tutorial recording policy, 17 were aware that there was one, 12 were not aware, and five were unsure. Only eight had read the policy. More broadly, most students (14) were not aware of how the university processes their personal data, 10 were unsure, and 10 were aware. Most students (22) were not aware of how Microsoft Teams processed their data.

44

When asked about their online learning and tutorial recording experiences, most students (20) said that some of their tutorials were recorded. In our study, we found that 18 students said that they were asked to consent to recordings for all of their online tutorials, five said only some asked for consent, seven were not asked, and four were not sure. In considering personal experiences of online learning, 11 said that online learning made a positive impact on their educational experience, two had no impact, 12 were impacted negatively, and nine were unsure. Focusing on tutorial recordings, 13 felt that tutorial recordings were a net positive, 17 did not feel that it impacted their educational experience, two were negatively impacted, and two were unsure.

Figure 8 shows the overall level of privacy concern of our participants, based on their responses to online learning IUIPC questions (Appendix C). The higher the score, the more privacy-concerned a participant is, where 55 is the maximum score and 11 is the minimum. Existing work shows that internet use reduces IUIPC. [79] A positive relationship was found between privacy concerns and government involvement in privacy regulation, [80] suggesting higher IUIPC scores for participants governed by the GDPR. The median score for our participants is 46. While our participants were based in the UK which falls under the GDPR’s remit students as young people are considered to have high levels of internet use, suggesting a relatively high level of privacy concern for their demographic. In assessing the significance of specific IUIPC questions for influencing a participant’s privacy concerns, from our exploratory factor analysis (TLI of factoring reliability = 1, RMSEA index = 0, and a confidence level of 95%), we found that for data collection, participants who thought about whether they should provide personal information to universities demonstrated higher levels of privacy concern, with a correlation of 0.8. Examining the IUIPC data awareness factor, the more important participants thought it was to be aware and knowledgeable about how their personal information will be used, the higher their IUIPC score, with a correlation of 0.9. From our analysis, consent as a form of privacy control was not significant enough to be considered as a factor for assessing the level of privacy concern.

Figure 8: The IUIPC scores of study participants. The median score for our participants is 46, demonstrating a moderately high level of privacy concern, where 55 is the maximum score and 11 is the minimum.

45

Figure 9 shows that most (28) participants consented to tutorial recording. We also asked participants to state whether they decided to change how they voted as a result of doing the exit survey. One participant from the commons group and two participants from the control group would change the way they voted based on the exit survey. All three changed from not consenting to giving consent. Most students (18) stated that they did not think twice before handing over their data to the university. This suggests that students may feel obliged to provide such data in order to access education and indicates a certain level of trust that students have of HEIs to use that data for academic purposes. Several participants across both groups stated that disability and accessibility were important reasons as to why they consented to the tutorial recording. In context of the COVID-19 pandemic when the study took place, this is particularly important given the challenges students face during online learning. As a result, it is necessary to consider accessibility needs when considering whether and how tutorial recording should be conducted to support students.

Figure 9: Consent preferences from participants answering the question “Should we record this tutorial?”. The majority of participants in both commons and control groups consented to tutorial recording.

46

Table 1 shows that resources related to data protection rights and policies were the most useful to our study participants. This suggests that participants are keen to increase their understanding of what protections are in place for their data and what recourse they have if anything goes wrong.

47

Interestingly, the actual consent voting poll where participants had to consent or not consent to the tutorial recording was the second least useful. This is consistent with existing literature on the limitations of meaningful and informed consent within and outside education that we outlined in Section 2. This demonstrates the importance for students to feel that they are making informed choices (where the outcome is less important) in an online learning environment, where they do not necessarily question the university’s motivations for recording tutorials. Additionally, given that the resources that give students more transparency were found to be marginally more useful that those that increased agency such as exercising rights and the quiz, more research could be done to examine how students believe their agency over their personal data could be increased.

Table 1: The resources in the commons that commons group participants found useful for helping them decide whether or not they should consent to tutorial recording. All participants found at least one commons resource to be useful.

48

When elaborating on why participants found certain features of the commons useful, one student said that “I hadn’t really known anything about tutorial recording policy or the laws and my rights related to these recordings before so I though (sic) it was interesting to learn more about my rights and more about what tutorial recordings would be used for and when they should be used”. Another student thought that “the information about the University policy was very valuable to make might (sic) decision, and having access to it easily is helpful. The FAQs was (sic) definitely the most helpful element, as it answered a lot of my questions simply”. A student who found the forum useful explained that: “I think I was most swayed by the anonymous student posts. Personally, I didn’t want the session recorded, but I knew it would be helpful for others to review later or who had missed the session/not been mentally present due to chronic illness, etc.”. Overall, nine participants in the commons group agreed that they would use the commons to improve the protection of their personal data. Five somewhat agreed, two neither agreed or disagreed, and one somewhat disagreed.

50

When conducting the study, we asked participants to imagine that they were taking part in a mock-tutorial on conducting research on social media and provided them with the lesson plan. During their post-study survey, we asked participants whether the topic of the mock-tutorial impacted their consent levels to tutorial recording (Figure 10 ).

Figure 10: Impact of the tutorial topic on consenting to tutorial recording. The topic of the tutorial is not a strong factor for informing whether a student decides to consent or not consent to the tutorial recording in both commons and control groups. The median for both the commons and control group is two.

51

From the survey responses, participants suggest that they would not refuse consent based on the tutorial topic alone as it would depend on other factors such as if they felt they needed to re-watch a tutorial recording and whether the topic involves providing personal information that the participants themselves did not want to share.

52

We wanted to increase our understanding of how students participate in recorded digital classrooms and how that may impact personal information sharing. To examine this, we asked participants whether they would avoid any topics during online learning, specifically those listed as special category personal data under GDPR Article 9. From Figures 11 and 12, the high number of avoided topics during tutorial recordings suggest that even if participants consented to tutorial recordings, teaching subjects that result in the discussion of these sensitive personal data may limit student participation in online learning. Two of the highest ranked topics ‘data concerning a person’ (22) and ‘political opinions’ (20) represent a broad range of information often shared in discussions. Six commons and three control group participants did not avoid any topic. Importantly, commons participants avoided fewer topics across all categories. This suggests that because commons participants have a more comprehensive understanding of how their data are stored, they are more comfortable having discussions about matters related to the special category personal data recorded. More generally, in examining the impact of the tutorial topic and the content participants are willing to share, they explained that they would rather not participate than not consent to the tutorial recording because they had control over what they said. As a result, it is important for staff to consider how to engage with teaching sensitive topics online to maximise participation and generate the most value from online learning.

Figure 11: Topics participants avoided in a recorded online learning environment from participants answering the question “Are there topics you will avoid discussing or revealing about yourself if the tutorial is recorded compared to physical classes?”. The two “other” responses include information that one participant considered to be “triggering” such as “mental health, other personal information, and financial information” as well as what another participant considered “anything that could be misconstrued or used out of context if the recording was inadvertently (or deliberately) released”. Overall, the commons participants are less likely to avoid discussing certain topics.

53

In response to our research questions, we find that interacting with commons resources helps inform students about the purpose of online learning and tutorial recordings. From our findings across both groups, students found the commons useful in supporting their data protection preferences. Some students in the control group would also have liked more information about how their data were being collected and used when being recorded.

54

The commons model is useful for supporting user preferences for protecting their personal data because it helps students develop a more detailed understanding of how their data are collected, used, and stored. Almost all participants consented to tutorial recording, indicating that students find value, both for themselves and for the class overall, when it comes to being able to access a recording if needed. Most students also indicated factors such as accessibility and helping other students as reasons for consenting to recordings. While recording the level of consent can be useful in understanding whether participants feel comfortable with the collection of data for tutorial recordings, it only provides a limited picture as to the extent to which the ability to interact with the commons helps inform them about the purposes of online learning and tutorial recording. Therefore, it is important to consider other factors and means for understanding student agency when it comes to supporting their data protection preferences. Our approach in creating a commons for data protection also addresses the limits of the individualistic approach in the GDPR as we move away from using consent from understanding how tutorial recording and online learning impact students’ education experiences.

55

More transparency around data protection between students, staff, university management, and other involved stakeholders is encouraged through a commons as the model supports the identification of stakeholder tensions and breaks them down through identifying a common aim: accessing a dynamic, participatory learning environment. The high levels of consent to tutorial recording could imply that students, to some extent, trust universities with their online learning data. However, given that more information on data protection was found to be preferable, students may want more details on how and what data are collected and used. This is particularly true in preventing potential harm should there be any data breaches, given the higher preference for understanding data protection and data subject rights. The commons model encourages more transparency around data protection between students and other stakeholders because it not only informs students of the data governance and risk management polices related to online learning data, but also supports recourse through data protection rights if any harms are realised.

56

More broadly, the commons model can be useful for reconsidering online learning pedagogies to support more inclusive and safe digital classrooms. Our results indicate that when asked about students’ participation in recorded online tutorials compared to in-person sessions, most participants indicated that there would be topics that they would avoid discussing. This suggests that while students are happy to consent to tutorial recordings, they may decrease their level of participation in online classes. This could impact the quality of tutorial participation in online teaching. As a result, staff should be mindful of asking students questions related to their personal experience that may reveal these forms of data. Commons participants are also more willing to reveal their personal data, suggesting that an improved understanding of what and how data are collected and processed can encourage participation. Overall, staff and academic institutions should consider how the online learning environment could be fostered to maintain the privacy and security offered by the physical classroom.

59

Our study has a number of limitations. Those who opted to participate are likely to be more privacy aware. Several participants mentioned that because data protection and tutorial recordings were mentioned in the study description, the thought was already on their mind beforehand. For unknown reasons, more students who identified as female participated in our study. Although we did not find any patterns or correlation to gender in response to our surveys, greater gender balance may be preferred to mitigate any potential biases.

60

Additionally, it was initially hoped that the study could have been done in groups to more accurately mimic the tutorial environment. However, challenges in recruiting participants, time zone difference, and asking them to spend more time on Microsoft Teams outside of classes meant that it was difficult to schedule participants to the same session. As a result, 28 students participated in the study individually and three pairs participated together. There was no identifiable difference between their responses.

61

Finally, we acknowledge that online learning during the pandemic is different to what it might have been if technologies were implemented more organically. Students, staff, and universities have had to instantly adapt to shifting the physical classroom into a digital one. The aim of our study is not to criticise HEIs for deploying technological solutions, but rather to encourage continued discussions on data protection considerations in education and suggest new socio-technical solutions that can help employ inclusive and innovative learning pedagogies to support the academic community.

62

According to the UK Higher Education Statistics Agency, there were 2.5 million students in higher education in the 2019/20 academic year, [81] where the vast majority of those students would have been online learning data subjects due to the pandemic. Despite the significant number, there is little attention given to improve support for how data in this sector should be used and protected.

63

While online learning safeguards are in place, as identified in Section many legal instruments and policies that can be applied to education are either general data protection principles or refer only to children’s data. Although these are useful and should be followed by HEIs, they do not adequately deal with the complexities of online learning data that can impact students beyond education. In the UK, this was exemplified with the UK Office of Qualifications and Examinations Regulation (Ofqual) A-Levels grading algorithm scandal, where the regulatory body used an equation to calculate secondary school students’ A-Levels results and therefore determined whether students were able to meet their university offers. [82] As the calculation heavily relied on the school’s historical predicted grades and grade distribution, many pupils felt that their algorithmic result did not reflect their examination abilities, with long-term consequences that not only affected their higher education but subsequent careers. Students in higher education have also begun to fight back. Students from the University of Amsterdam, [83] University of

64

Maastricht, [84] City University of New York, [85] and University of Texas, [86] amongst many others, have organised petitions to push back against e-proctoring technologies not only because of privacy violations but also the technology’s discriminatory nature and for fostering a surveillance-based academic environment. From our study, it is notable that students do care about what happens with their data, even if they trust institutions with it. As a result, in addition to the responsibility of providing education, HEIs should also have ethical responsibility to ensure that the use of student data by institutions themselves and third parties are actively communicated. This cannot be done through data protection alone and must include broader considerations of digital infrastructures and data governance strategies within higher education.

65

More broadly, wider conversations between data protection stakeholders could be facilitated to raise awareness on how personal data is being treated in our data-driven society within a data-protection focused data commons from legal, socio-technological, and ethical perspectives. This includes discussing the impact of data-related regulations and policies on data subjects. For example, within data protection regulations, access to the fundamental right to data protection through the exercise of data rights can be further strengthened. [87] Laws such as the European Data Governance Act [88] and Data Act [89] aims to increase trust in data intermediaries and strengthen data-sharing mechanisms across the EU, could support broader data protection practices for the benefit of data subjects outside of data protection. In the UK, from a group privacy perspective, Lloyd v. Google [90] raises interesting questions on collective action for privacy violation claims, which may lower the bar for collective redress actions. [91] Research and guidance from organisations and advisory bodies such as the Centre for Data Ethics and Innovation in the UK can play an important role connecting different stakeholders and addressing data issues to specific domains, including the data infrastructures needed to support a commons. [92] As a result, legal developments fostering the use of collaborative and co-created data-related practices can support greater fairness, accountability, and transparency on how data can be used for the benefit of individuals and groups. As students are becoming more aware of how online learning technologies use their personal data in and outside of classrooms, they have increasingly put pressure on HEIs to look beyond data protection considerations when deciding to adopt such technologies, focusing on the ethical, health, and wellbeing aspects of using educational data. Reiterating the limitations of consent from our study findings in Section 6, it is important that matters beyond lawful basis of data processing are considered when it comes to how students can have agency over their online learning data and experience. This requires university management, data protection authorities, regulators, and policy makers to consider new ways in which online learning data should be regulated and governed to protect student data while also generating value for education.

66

Given that our online learning commons was only tested on students, further research could be done with staff to examine whether the commons could be useful for protecting their agency for protecting personal data. This is particularly important due to concerns of HEIs using educational technologies to monitor staff [93]and break union strikes, [94] where intellectual property rights do not always belong to the individual who produced the work. [95]  [96] With the rise of children’s data collection in the classroom, the commons could also be tested on younger learners to examine its usefulness for students, teachers, and parents.

67

As the commons aims to provide a socio-technical framework for increased protection and for expressing data protection preferences, it can be applied to other use cases both within online learning as well as other domains. For example, a data protection-focused data commons could be adopted for social media data archiving, where individuals may have preferences as to what, how, to whom, and for how long their posts are shared. Those who are part of the commons can also find out if there have been any recent scandals or data breaches related to different platforms. As the data protection-focused commons can be useful for demystifying the data protection regulations, policies, and processes for data subjects, future work can help identify new use cases for the benefit of communities across different sectors. Crucially, the application of the commons is not exclusive to countries that fall within the GDPR but can be tested in other jurisdictions to support a socio-technical means for greater student agency and data protection.

68

Beyond the commons, teaching pedagogies should be re-examined to ensure that students and staff are able to make the most of online learning technologies without losing some of the privacy and security offered by the in-person learning. More research could be done to examine whether student attitudes towards online learning has changed as they return to the in-person classroom. As students no longer have to fully rely on technology for academic study and are able to compare the online environment to in-person teaching, it can be expected that they are able to more accurately identify the benefits and downfalls of both.

69

More generally, given the rapid deployment of technologies to enable online learning delivery during the pandemic, there was understandably little time to explore how digital tools could be used to enhance learning within online education spaces. As a result, this period of digitisation may be characterised as a missed opportunity for co-creating teaching methods, forms, and pedagogies to support creative educational practices that extend and supplement physical classroom activities. To limit only replicating in-person educational norms using digital tools, as the impact of the pandemic eases, inclusive efforts to re-imagine the purpose of education and the role of technology through socio-technical lenses and education-related policies may help develop collaborative educational practices to support digital learning.

70

Considering the future development of the commons, to deploy a data commons in the long-term, considerations need to be made with regards to the platform used to host the commons and how the commons is to be sustained financially. While our study used Microsoft Teams, it does not represent the ideal platform for the commons. When deciding where and how the commons should be hosted, commons stakeholders should be involved in the decision-making process. Further work in this area could include developments on technical infrastructure and system considerations related to personal data, particularly whether the commons could be created within an existing digital ecosystem or built independently. Development decisions should be made in consultation with data subjects based on their accessibility, data, and data protection requirements, as well as expert advice to ensure that adequate checks and balances are in place to protect data that is processed within the commons. Given the difference in stakeholder interests, how and by whom the commons is maintained can impact the trust between users as data subjects and others participating in the commons’ development.