Call for a new data governance structure

A summary of the new book chapter for the DFC Book Collection Education Data Futures.

by: Jun Zhao

 
21 Nov 2022

A huge amount and range of education data is routinely collected in UK state schools (Persson, 2020). These data can be collected and processed at schools for a variety of purposes. Sometimes, schools are collecting these data under their obligation to the Department for Education (DfE); other times, they need to process children’s data as part of teaching, assessment, administrative and safeguarding (DFC, 2021b). Finally, schools are increasingly contracting external EdTech companies to process children’s data to enhance their learning and education opportunities.

While schools are expected to be the primary duty bearers of children’s best interests, their duties are compounded by the complexity of legislation in the education sector and the extreme challenge of carrying out compliance validation. To shed lights upon this challeging situation, the Digital Future Commissions and 5Rights Foundations called for a timely peer-reviewed essay collection, namely Education Data Futures, which was launched on 21 November 2022.

Researchers from the EWADA project contributed our thoughts of a technical alternative to the current learning data management and governance paradigm for UK state schools. Our proposal is based on the new wave of decentralised paradigms for data sharing and ownership. This decentralised data governance is anticipated to expand individual data subjects’ ability to access data and establish data autonomy. A data trust is one such data governance model, which provides a promising response to schools’ need for an independent and trustworthy body of experts who can make critical decisions about who has access to data and under what conditions.

A data trust as ‘a legal structure that provides independent stewardship of data’, including deciding who has access to data, under what conditions and to whose benefit (Open Data Institute, 2019). It is different from other data governance structures because it represents ‘a legal relationship where a trustee stewards data rights in the sole interests of a beneficiary or a group of beneficiaries’ (van Geuns & Brandusescu, 2020). Instead of taking a grassroots governance model (such as a data commons), the trustees can be the decision-makers regarding who has access, under what conditions and to whose benefit, and they take on a legally binding responsibility for data stewardship (Open Data Institute, 2019).

It has been exciting to see some practical developments of data trusts recently, built on extensive theoretical landscaping. However, developing data trusts is a complex task and requires a strong commitment from data holders and the users’ community. Furthermore, existing legal frameworks are not necessarily ready to support all the data stewardship and legal binding responsibilities designed for a data trust. The essay runs through a case study about personalised learning at school to demonstrate what a data trust model may provide.

As shown in the figure below, a data trust can act as an intermediary for schools, making decisions about what education data can be accessed by a third party, investigating the purposes of data access, and assessing how they may be aligned with students’ best interests. It provides a promising direction for mitigating the challenges that schools face regarding data safeguarding and compliance checking.

However, developing such a data trust is not without challenges. The essay concludes by discussing open social, legal and technological challenges to be considered, calling for a pilot model of data trusts in the educational technology (EdTech) sector, for which the technologies of the EWADA project could notable contributions to.

Education data trust

Please follow the link for further reading about the essay pdf and the essay collection.

EWADA Summer 2022 Internship Report

A summary of the four projects carried out and how they contribute to the EWADA vision.

by: Jun Zhao

 
31 Oct 2022

Summer 2022 is our second year of running a very successful internship programme. We are fortunate to have secured four internships with excellent candidates, each adding tremendous contributions to EWADA’s exploration of creating ethical web applications for the future with Solid. This summer has been full of excitement and inspiration. This report summarises the four projects carried out and how they contribute to the EWADA vision.

Overview of the projects

The key goal of the EWADA project is “to develop and deploy new forms of technical and legal infrastructure, to re-design and promote a more equitable and ethical treatment of individual users and collectives in a sustainable way”. In the first year, our focus is to deliver a suit of case study applications so that we could gain a deeper understanding of the components that are needed for building the ethical, decentralised architecture. For this, we need to understand both from a user’s point of view and a practical and technical point of view what these components should look like.

From a human-centred point of view, we need to understand what users recognise as being ethical, how they would like these components to look like and what supports are needed. For the engineers and researchers, we need to understand what technical are needed and how they could be brought together.

We have gained some knowledge about these challenges through our ongoing research with children (regarding datafication and data autonomy) and our building of a solid-media application. This summer, we aim to achieve further understandings through the following projects:

  • The Gigworker project: by interacting with gig workers to explore what data autonomy barriers that workers are facing with existing online platforms and what support they need
  • The Solid Calendar project: by extending existing Solid architecture’s capability of interoperating with existing data sources, s uch as various calendar formats from existing providers
  • The SolidFlix project: by benchmarking performance of Solid-based applications and exploring ethical movie recommendation opportunities
  • The Multi-Party Computation (MPC) project: by extending existing Solid-like decentralised systems with the necessary privacy-preserving computation capabilities.

As shown in the outline figure below, these projects provide a rich opportunity to explore various technical and social components for realising our vision of an ethical architecture:

  1. Supporting collective users’ values: Existing solid-based applications have largely been focusing on supporting individual users to explore and manage their own data using Solid, in order to bootstrap their autonomy with their data. However, the ‘GigWorker’ project aims to explore how Solid can support collective values from individuals’ autonomy, for example for workers to argue for more fair salary rate or to create peer support, reduce isolations and manage their collective well-beings. For this, we are interested to examine what kind of collective values would be useful for workers, and what concerns and needs they may have.

  2. Supporting a richer decentralised web architecture: Access control is fundamental to Solid data storage. However, several additional components may be useful in addition to the vanilla Solid. This includes: 1) a privacy-preservation computation component, so that we could apply data aggregations or AI-based computations to individual’s data without compromising their data privacy and security; 2) an ethical AI computation component, so that personalisation computations will not continue to introduce filter bubbles, biased recommendations or exploitations; and 3) a data adapter/orchestrator component, so that Solid applications can interoperate with existing data sources or formats, without forcing users to abandon their existing digital habits.

EWADA Summer 2022
Fig 1. EWADA summer projects contribute to new ethcial app components (a Data Orchestrator, an ethical AI recommender and a privacy computation component), and three new ethical (social) web applications, including a social calendar, an ethical platform for gigworkers and a platform for social moving sharing.

Outcomes and next steps

Our summer projects made great strides in exploring these social and technical components. From the GigWorker study, we have identified the plurality of needs from the workers and we are continuing the work by creating a prototype that will cater for workers’ diversity needs. We generated encouraging early results for integrating a privacy-preserving computation components with the Solid architecture, and we aim to extend this work by deploying this integration in actual ethical applications. We completed the prototyping of the data orchestrator design and are hoping to present this design to the Solid community in the near future. In the process of adding ethical recommendations to the SoliFlix project, we worked extensively to explore performance optimisations and we are also hoping to present this result to the Solid community soon.

Fig 2 illustrates our key next steps:

  • From in principle to in practice: several components mentioned above have not yet been fully integrated into one reference implementation, which will present an exciting starting point for developing practical application scenarios such as for gig workers or ethical recommendation systems for children.
  • From in principle to in the wild: we will also focus on sharing all the design and development experiences with the wider Solid and open-source communities, to ensure interoperability and bootstrap the uptake of our proposed solutions.
  • Furthermore, we will consolidate some of the components we started and make investigation of various social-technical components, based on this excellent technical ground we have developed so far, including how we may provide supports for users to select Solid pod providers or ethical app components that they can trust, what kind of algorithmic transparency will be needed by users in the newly decentralised and ethical web applications, what new data protection regulations, governance and auditing are needed, and etc.

We are hoping the successful summer internship will continue its tradition next year as it has been exciting to see how the internships have provided an effective way for us to define miniature tasks that make critical contributions to the project overall, as well as to transfer our frontier understandings about the key of creating web decentralisations to our amazing interns, Vid, Zimeng, Hunar and Adrien! Thank you, and wish you all the best of luck!

Post EWADA Summer 2022
Fig 2. EWADA's focus post-summer internships: deployment and testing of new applications, and enhancement of core ethical computing components.

Welcome a new EWADA DPhil

Welcome a new EWADA DPhil

by: Jun Zhao

 
01 Oct 2022

We are really excited to welcome our new full-time EWADA DPhil joining the project - Zhilin Zhang.

Zhilin received his B.S. and M.S., both in Computer Science, from the University of Illinois at Urbana-Champaign (UIUC). In EWADA, Zhilin will study HCI to better understand and improve human interactions with AI and other complex algorithmic systems. I design and build computing systems to positively support users’ online behaviors and interactions in a scalable and accessible way.

Zhilin is fully-funded by the prestigious Oxford Computer Science Departmental Studentship and co-supervised by Professor Nigel Shadbolt and Dr Jun Zhao.

EWADA project summer internships

EWADA project offers three summer internship positions in 2022

by: Jun Zhao

 
17 Jun 2022

We are very excited to announce 3 summer internships in 2022!!!

Please come to join us if you want to create an AUTONOMOUS Web with a group of world-leading computer scientists!

Detailed job description

You will be working with researchers from [Oxford HCC (http://hcc.cs.ox.ac.uk))] to develop new web applications that enable web users to take back the control of their personal online data and decide how an application may make use of their data in a privacy-preserving way.

You will be working with the Solid (Social Linked Data) architecture from MIT [1], which is a standards-based approach that provides practical solutions to web decentralisation and empowers people to take back control of their data and regain their self-autonomy.

Background of the project

You will be working as part of the Oxford Martin School programme EWADA (Ethical Web and Data Infrastructure in the Age of AI)[2] Project, which is an ambitious 3-year programme that aims to reform the concentration of power on the Web by developing and deploying new forms of technical and legal infrastructure. The project is led by Prof. Sir. Nigel Shadbolt and Prof. Sir. Tim Berner-Lee and aims to investigate novel re-decentralisation architectures and develop critical privacy-preserving AI methods to re-establish citizens’ self-autonomy on the Web.

Selection criteria

Ideally, you are expected to have experience with full-stack web development and be interested and willing to learn about EWADA research and Solid ecosystems (in Javascript or Python). Moreover, you will need to be able to work well with others, as these roles will be highly collaborative.

We have a range of application choices for this year’s summer internship, including:

  • Creating a solid-based decentralise social media platform so that we do not have to be confined by existing proprietary solutions
  • Creating a solid-based worker platform that enables gig workers (such as Amazon workers or Uber drivers) to take better control of their data
  • Implementing a privacy-preserving computational protocol for users to share their data in a solid-based decentralised architecture

Application

The post is expected to be full-time (36.5 hours) for 12 weeks, starting mid-July 2022 and ending in September 2022, £12.74-14.29 (Grade 3.8-4.7) per hour, depending on experience. If you are a student holding a Tier 4 visa, then you are permitted to work full-time for 8 weeks and 4 weeks part-time (max 20 hrs per week).

The post does not have to be based in Oxford but will be subject to the right to work in the UK. We CAN NOT sponsor visa applications due to the short duration of the project.

Applications should be submitted to Dr Jun Zhao (jun.zhao@cs.ox.ac.uk), including a 2-page resume, an indication of the application of interest, and a short paragraph of background and interests.

The closing date for applications is 12 noon on Friday, 1 July 2022. Candidates will be shortlisted and invited for an interview by 8 July.

Selection criteria

Essential

  • The ability to communicate information clearly, including technical content
  • The ability and desire to learn about Solid-related research, to quickly acquire domain expertise needed for effectively developing new web applications
  • The ability and desire to work with research staff to elicit information about how information should be organised
  • Substantial experience developing web-based applications
  • The ability to effectively manage time, to complete projects efficiently
  • The ability to work independently and think creatively

Desirable

  • Expertise with base Web platform languages & technologies including HTML, CSS and Javascript
  • Expertise with Python and Linux-based Operating Systems
  • Graphic and interaction design experience and expertise

EWADA first year

EWADA first year project meeting

by: Jun Zhao

 
28 Mar 2022

On 26-27 March, EWADA researchers attended the Venice Colloquium 2022, co-organised by Sir Nigel Shadbolt, the Royal Society and Goodenough College. The colloqium assembled researchers from Oxford and from Goodenough working in AI, philosophy, legal regulation and public policy to consider how best the development of AI might be framed and regulated across different governments and societies. The Colloquium hopes to sponsor genuine cross-disciplinary exchange between different academic communities.

Oxford researchers contributed a series of talks about AI for children, Privacy and Data Protection in an age of AI, AI regulation for the use of social media, as well as Ethical Data Architectures. A series of essays are expected to be published later this year, reporting all detailed discussions taking place in the Colloquium. We very much look forward to a productive collaboration ahead!

Welcome new EWADA researchers

Welcome new EWADA researchers

by: Jun Zhao

 
07 Mar 2022

We are really excited to welcome two new full-time EWADA team members on board:

  • Dr Naman Noel, previously affiliated with ETH Zurich
  • Rui Zhao, a final PhD student from University of Edinburgh

Their further details can be found in our team page.

We cannot wait to further our exploration of the open and privacy-preserving architecture building with these two new talents!

EWADA response to CMA consultation

Oxford's response to CMA consultation

by: Jun Zhao

 
14 Feb 2022

Researchers from Oxford Human Centred Computing Research Group (of which EWADA is part of) provided a response to the public consultation on mobile ecosystems by the Competition and Markets Authority.

Mobile ecosystems serve as an important part of our everyday lives, mediating social, political and market interactions. The ever increasing importance and ubiquity of mobile devices puts great power into the hands of those companies that make the key design decisions affecting mobile ecosystems. Decisions that might sometimes go against the interests of consumers. This is why we welcome the opportunity to respond to the ongoing investigation of the CMA into mobile ecosystems. Oxford HCC/EWADA researchers have been studying the data flows and power relations in the mobile ecosystem for many years, and thus welcome the opportunity to respond to the ongoing investigation of the CMA into mobile ecosystems.

The major themes of our response cover the following topics:

  • Ensure that app developers are not (implicitly or explicitly) nudged into violating basic provisions of UK data protection law, particularly the need to seek consent before engaging in third-party tracking. This could include standards for regulatory conformance (e.g. clarifying the responsibility of those companies developing tracking technologies, and requiring them to provide simple and compliant implementation guidance to app developers), and should ultimately aim to build a mobile ecosystem that facilitates compliance by default (see Section 1.1.1).
  • Empower researchers to conduct app research, by enabling ways in which researchers can more easily analyse encrypted iOS apps, download apps at scale, and analyse encrypted network traffic of apps on Android (see Sections 1.2.1–1.2.2).
  • Enable researchers to analyse concerns around underlying technologies of the mobile ecosystem, including the use of data relating to individuals and other advertising companies in Apple’s SKAdNetwork (see Section 1.4.3). Lower barriers to entry and innovation. Encourage the use of cross-platform technologies in app development (such as open web technologies), ensure that Windows and Linux users can develop apps for iOS (currently only macOS users), and lower the barrier to entry into the App Store (currently an annual 99 USD fee applies) (see Section 2.1).
  • Ensure that gatekeepers do not self-preference, particularly with regards to ad attribution or in the definition of tracking in the Apple ecosystem, and the distribution of adblocking technologies on Android and in Google Chrome (see Sections 1.4.2–1.4.3 and 2.1). Scrutinise Google’s current ban of in-app tracking blockers (see Section 2.1), including Disconnect.me, to give consumers more choice over how apps use their data and to tackle widespread infringements of data protection law (particularly the need to seek user consent prior to tracking, as well as proportionality, data minimisation and purpose limitation) within apps.
  • Consider requiring smartphone OS’s and app store operators to enable third-party mobile app extension functionality to spur innovation in mobile apps and reduce harms within them, similar to the approach taken with extensions in desktop browsers and mobile Safari starting with iOS 15, while ensuring safety of consumers using such extensions through the existing app store review processes (see Section 2.2).
  • Ensure that the review of apps on the app stores and the policies underlying this process are fair and transparent, for example through regular mandatory disclosures about this enforcement (including with regards to privacy and data protection). Such disclosures would be a minimally invasive but realistic intervention, and have been suggested by a variety of researchers from different backgrounds. See Section 3.1 for more details.
  • Consider separating key functions within the governance of mobile ecosystems to reduce conflicts of interests, such as privacy management to avoid self-preferencing as regards data collection and protect consumers against excesses and monopolisation of such data collection, and promoting more research into this area (see Section 3.2).

Read our full responses here

EWADA are supporting Leys-Oxford App Club

EWADA researchers are involved in the app club to support Leys CDI young people

by: Jun Zhao

 
31 Jan 2022

The Leys-Oxford App Club supports young people from the Leys Community Development Initiative who are passionate about learning new mobile app development skills and broadening their experiences and knowledge. It is part of Oxford University’s SCIENCE TOGETHER initiative, which is a brand new, grass-roots programme that harnesses the power of community-led collaborative research projects, bringing together Oxford scientists and people who live and work in Oxfordshire, to address major issues and questions impacting local people’s lives. In the case of App Club, it brings Oxford University computer scientists to the youth centre for closer knowledge sharing and co-creation opportunities.

The mission of the club is to transfer step-by-step app development skills and knowledge to club participants over a six-week period. The objectives are for the young people to learn about the life cycle of app development, understand the diverse skills involved in the app development, and develop team collaboration skills.

Through initial explorations, we have designed a 6-week app club programme, with a clear objective of producing an app prototype that can help young people to digitise the process of signing-in an event at the centre so that they can avoid the existing paper-based process and the centre can keep a digital record of the attendance. The young people are passionate about the potential of the app for simplifying their current experiences with the centre and helping with the promotion of the centre.

The outcome app ‘CDI app’ will be further tested with young people visiting the centre in the summer and jointly with the new Leys CDI website. We are very excited about this new initiative and a new form of ‘Science Together’.

EWADA response to DCMS consultation

Oxford's response to DCMS consultation

by: Jun Zhao

 
19 Nov 2021

Researchers from Oxford Human Centred Computing Research Group (of which EWADA is part of) and the Responsible Technology Institute provided a joint response to the public consultation on reforms to the UK’s data protection regime by the Department for Digital, Culture, Media & Sport (DCMS).

The goal of the consultation is to create an ambitious, pro-growth and innovation-friendly data protection regime in the UK that underpins the trustworthy use of data. UK data protection law is currently modelled on EU requirements, particularly the GDPR from 2016 and the 2009 ePrivacy Directive. Following the UK’s withdrawal from the European Union, the government sees a wealth of new opportunities in the reform of the current legal requirements around the protection of personal data.

Overall, as a research group, we welcome the initiative of the UK government to help researchers in handling personal data, and to spur innovation within the UK. At the same time, we are concerned about potentially harmful consequences for individuals residing within the UK, as a result of a potential weakening of UK data protection standards and rights. Our response highlighted our thoughts around the core issues of data protection and privacy, responsible research, and balancing support for organisations as well as individuals. The major themes of our response cover the following topics:

  • Data intermediaries and institutions: Lack of clarity regarding data intermediaries, institutions, and practices put in place to safeguard individuals and support technological growth.
  • AI and responsible innovation: The opportunities for AI innovation in the UK depend on a robust regulatory regime that encourages highly context-specific risk management. This will be best promoted through maintaining existing measures like Data Protection Impact Assessments, Data Protection Officers, record keeping, and prior consultation, amongst others.
  • Erosion of trust in online tracking: Excessive box-ticking in the form of consent banners is not a necessary feature of existing data protection and privacy law, but rather a symptom of non-compliance with it.
  • Removal of the balancing test: The removal of the balancing test for pre-approved legitimate interest purposes will create disproportionate risks for UK citizens, and a false sense of certainty for controllers.

In our response, we also provided recommendations and suggestions on these themes and statements, intended to help build a sustainable future of AI and data protection within the UK that not only promotes innovation but also advocates for and protects individuals.

Read our full responses here

Welcome Research Visit from Prof Ruben Verborgh

We hosted first research visit from Prof Ruben Verborgh from Ghent University

by: Jun Zhao

 
17 Nov 2021

On 17-18 November 2021, the Oxford EWADA team welcomed the visit from Professor Ruben Verborgh from Ghent University, who is a world-leading expert in decentralized Web technologies.

The initial introduction meeting was well attended by the core EWADA CS team as well as many graduate and undergraduate students from the Oxford HCC group.

We discussed the latest development of the Solid Community Server led by the Ghent iMind research centre, Inrupt’s strategy and EWADA’s progress so far. The research visit has led to a productive identification of the common roadmaps between the Oxford and Ghent teams and a set of agreed forthcoming developments, which will undoubtedly strengthen our knowledge exchange and the delivery of the EWADA vision in a broader and impactful way.