We are super proud to share that Sir Tim Berners-Lee’s new book – This Is for Everyone: The Unfinished Story of the World Wide Web – has now arrived at Oxford Blackwell (Waterstone), ready to order. Arriving at a crucial moment, the book is more than an autobiography, but a call to re-imagine the architecture, ethics, and values underlying our digital lives. In the book, Tim has not just recounted how he invented the Web, but he challenges us to reclaim it.

Today, the Web is becoming one of most powerful digital infrastructures in the world. However, Tim and our whole research team have become so deeply concerned about how platforms are invasively harvesting personal data, enabling the rapid widespread of misinformation and personalised behaviour manipulation. Tim’s book provides the much-needed discussions about how the Web’s infrastructure and governance have allowed commercial and political forces to override many of its original democratic, creative roots” Washington Post.

With the rapid advances in AI and associated centralisation of power in tech platforms, with weakening privacy protections, the stakes are higher than ever. Tim’s book contends that unless we rethink not just what the Web does, but how it’s structured (its protocols, its governance, its norms), its promise may slip away.

As the co-PI of the EWADA, Tim has devoted his valuable time guiding us towards this vision of his – re-enabling the openness of the Web for everyone. This is particularly exemplified by our work in creating a new protocol for enabling data autonomy through an expression of term of use, new architectural designs to enable privacy-preserving AI computations in the decentralised settings, and many of our research studies with children,gig workers, social media users and administrators, as well as broad general public, to understand what they need to have a better Web, better control and autonomy of their data.

We hope you will all find Tim’s new book an inspiring read, to join our force to reclaiming the Web for everyone!

We are pleased to share that our upcoming workshop of “Human-Centric Agentic Web (HAW)” has been accepted and will be co-located in London, at Decentralised AI 2025 on 21 November, 2025.

Recent advances in large language models (LLMs) have ushered in a new era of intelligent agent technologies. These agents, built on top of powerful foundation models, are showing promising performance in long-horizon tasks, collaborating with humans and other agents, and acting autonomously across digital environments. However, the current discourse surrounding LLM-based agents has been heavily focused on algorithmic innovation – prompt engineering, planning and reasoning mechanisms, memory architectures, and tool use — while comparatively little attention has been paid to the foundational systems and infrastructural underpinnings that will be critical for deploying these agents at scale, safely, and for the benefit of users.

This workshop aims to explore the emerging infrastructure required to support safe, user-centric, and decentralised AI agents. Unlike classical multi-agent systems that often operated in controlled, closed environments, LLM-based agents are poised to operate openly and widely across the internet, potentially interacting with other agents and humans across jurisdictions, platforms, and use cases. This introduces new challenges in identity management, communication protocols, access control, privacy, auditability, availability and quality of inference data, interoperability, and alignment with user intent. These are not merely engineering problems; they require careful rethinking of how we design agent systems that are robust, accountable, privacy-preserving, and work for diverse stakeholders.

This workshop will convene discussion around novel architectures, system design patterns, protocol development, data interoperability, data quality, decentralised governance models, human-in-the-loop safety mechanisms, and standards for inter-agent communication. By bringing together researchers from multi-agent systems, systems engineering, security, HCI, and AI ethics, the workshop seeks to chart a path toward responsible infrastructure for next-generation AI agents.

Workshop Organisation Commitee

Core Organising Team

• Dr Panayiotis Danassis, University of Southampton
• Dr Naman Goel, University of Oxford
• Jesse Wright, University of Oxford
• Dr An Zhang, University of Science and Technology of China

Steeing Committee

• Prof Sir Nigel Shadbolt
• Dr Jun Zhao
• Dr Rui Zhao

Workshop formats and duration

Further details about the workshop format and duration will be shared on the dedicated workshop page. Please look out for this space.

We are pleased to annouce that this summer EWADA is launching a new research study with children aged 13-15!

Ever wonder what determines what your teenager sees on TikTok, Instagram, or Netflix? An overview of our study can be found from the video below.

EWADA team have developed new special social media apps that can help children learn how their data is used by existing social media platforms, and how they can take better control of their data.

This summer, we’re looking for families with teens aged 13-16 to join our research study. Children will try out two fun and easy-to-use apps — CHAITok and CHAIFlix — designed to help children better understand and manage their personal data on social media.

• Who: Families with teens ages 13-16
• What: A research session exploring children’s personal data on social media platforms
• Why: To give teen insight into the invisible forces shaping young people’s digital experiences

To the join the study and help your child understand better the digital world they navigate every day, register your interest here.

You can learn more about the tools we will use in the study with the short video here.

We are pleased to share that EWADA researchers Dr Rui Zhao and Jesse Wright presented their paepr at the Ethical Web Science Workshop, co-located at the WebSci’25 in New Jersey, USA.

Their paper – Introduce an Auditing Layer to Web Science – presents the conceptual design of a novel framework for to ensure the capability of auditing in data-harvesting Web Science practices.

Abstract:

Scientific discoveries increasingly depend on data and data processing, and Web Science is no exception. As an established practice, data-intensive research typically uses scientific workflows and provenance to facilitate data and method sharing while automatically preserving processing history. Prior research has reported the possibility of ex-post policy-based compliance checking from provenance data. Based on these works, in this paper, we present the conceptual design of a framework of data-harvesting Web Science practices, especially by introducing a common auditing layer. We discuss the framework’s practical, scientific, and ethical advantages, including its applicability in the period of large language model (LLM), autonomous agent, and artificial intelligence (AI) explosion. We hope this framework design can incubate a new norm for research practice to be transparent, ethical, and lightweight.

We are pleased to share the EWADA team have a new full paper accepted by the premier AI conference – FaccT 2025.

Title: Not Even Nice Work If You Can Get It; A Longitudinal Study of Uber’s Algorithmic Pay and Pricing

Abstract:

Ride-sharing platforms like Uber market themselves as enabling ‘flexibility’ for their workforce, meaning that drivers are expected to anticipate when and where the algorithm will allocate them jobs, and how well remunerated those jobs will be. In this work we describe our process of participatory action research with drivers and trade union organisers, culminating in a participatory audit of Uber’s algorithmic pay and work allocation, before and after the introduction of dynamic pricing. Through longitudinal analysis of 1.5 million trips from 258 drivers in the UK, we find that after dynamic pricing, pay has decreased, Uber’s cut has increased, job allocation and pay is less predictable, inequality between drivers is increased, and drivers spend more time waiting for jobs. In addition to these findings, we provide methodological and theoretical contributions to algorithm auditing, gig work, and the emerging practice of worker data science.

We are pleased to share the EWADA team member Dr Naman Goel has a new demo paper accepted by the premier AI conference – IJCAI 2025. The paper is a collaoboration with researchers from University of Edinburgh, Jain Family Institute, Utrecht University, and Accenture.

Gala et al. Fairness-Aware Interactive Target Variable Definition (Demo). Proceedings of the 34th International Joint Conference on Artificial Intelligence, IJCAI, 2025. Paper; Demo

Title: Fairness-Aware Interactive Target Variable Definition

Abstract:

Machine learning requires defining one’s target variable for predictions or decisions, a process that can have profound implications on fairness, since biases are often encoded in target variable definition itself, before any data collection or training. The downstream impacts of target variable definitions must be taken into account in order to responsibly develop, deploy, and use the algorithmic systems. We propose FairTargetSim (FTS), an interactive and simulations-based approach for this. We demonstrate FTS using the example of algorithmic hiring, grounded in real-world data and user-defined target variables. FTS is open-source; it can be used by algorithm developers, non-technical stakeholders, researchers, and educators in a number of ways. FTS is available at: http://tinyurl.com/ftsinterface. The video accompanying this paper is here: http://tinyurl.com/ijcaifts.

Position paper accepted the ICML

We are pleased to share the EWADA team member Dr Rui Zhao has a new paper accepted by the premier AI conference – ICML 2025. The paper is our first collaoboration with researchers from National University of Singapore.

Moming Duan, Mingzhe Du, Rui Zhao, Mengying Wang, Yinghui Wu, Nigel Shadbolt, Bingsheng He. Current Model Licensing Practices are Dragging Us into a Quagmire of Legal Noncompliance. The 42nd International Conference on Machine Learning (ICML 2025)

Title: Current Model Licensing Practices are Dragging Us into a Quagmire of Legal Noncompliance

Abstract:

The Machine Learning (ML) community has witnessed explosive growth, with millions of ML models being published on the Web. Reusing ML model components has been prevalent nowadays. Developers are often required to choose a license to publish and govern the use of their models. Popular options include Apache-2.0, OpenRAIL (Responsible AI Licenses), Creative Commons Licenses (CCs), Llama2, and GPL-3.0. Currently, no standard or widely accepted best practices exist for model licensing. But does this lack of standardization lead to undesired consequences? Our answer is Yes. After reviewing the clauses of the most widely adopted licenses, we take the position that current model licensing practices are dragging us into a quagmire of legal noncompliance. To support this view, we explore the cur- rent practices in model licensing and highlight the differences between various model licenses. We then identify potential legal risks associated with these licenses and demonstrate these risks using examples from real-world repositories on Hugging Face. To foster a more standardized future for model licensing, we also propose a new draft of model licenses, ModelGo Licenses (MGLs), to address these challenges and promote better compliance. https://www.modelgo.li/

The paper can be found at pdf and the presentation can be found here

We are pleased to share that our paper “Libertas: Privacy-Preserving Collaborative Computation for Decentralised Personal Data Stores” has been accepted by the preminer HCI conferfence CSCW, to be presented in November 2025.

Data and data processing have become an indispensable aspect for our society. Insights drawn from collective data make invaluable contribution to scientific, societal and communal research and business. But there are increasing worries about privacy issues and data misuse. This has prompted the emergence of decentralised personal data stores (PDS) like Solid that provide individuals more control over their personal data. However, existing PDS frameworks face challenges in ensuring data privacy when performing collective computations that combine data from multiple users. While Secure Multi-Party Computation (MPC) offers input secrecy protection during collective computation without relying on any single party, issues emerge when directly applying MPC in the context of PDS, particularly due to key factors like autonomy and decentralisation. In this work, we discuss the essence of this issue, identify a potential solution, and introduce a modular system architecture, Libertas, to integrate MPC with PDS like Solid, without requiring protocol-level changes. We introduce a paradigm shift from an ‘omniscient’ view to individual-based, user-centric view of trust and security, and discuss the threat model of Libertas. Two realistic use cases for collaborative data processing are used for evaluation, both for technical feasibility and empirical benchmark, highlighting its effectiveness in empowering gig workers and generating differentially private synthetic data. The results of our experiments underscore Libertas’ linear scalability and provide valuable insights into compute optimisations, thereby advancing the state-of-the-art in privacy-preserving data processing practices. By offering practical solutions for maintaining both individual autonomy and privacy in collaborative data processing environments, Libertas contributes significantly to the ongoing discourse on privacy protection in data-driven decision-making contexts.

The full paper can be found arxiv with the code accessible from Github

On Feb 26, 2025, Prof Ruben Verborgh, Sir Tim Berners-Lee & Sir Nigel Shadbolt gave a public seminar on ‘Towards re-decentralising the Web: an ethical web and data infrastructure’ at the Oxford Martin School.

In this talk, Prof Ruben Verborgh, Visiting Fellow of EWADA, was joined by Prof Sir Tim Berners-Lee and Prof Sir Nigel Shadbolt and discussed how new forms of technical and legal infrastructure are being developed and deployed to provide a more equitable and ethical treatment of individual users in the age of AI. The seminar outlined the challenges of achieving better data autonomy, providing individuals with better rights to data privacy while enabling them to benefit from the exciting possibilities of digital innovations.

The seminar, chaired by Professor Sir Charles Godfray, was well-attended by students, researchers from the university, and members of the general public. It began with an insightful overview of the need for a new form of data governance, followed by an introduction to the Solid architecture and a review of current developments in the deployment of decentralized data systems. The session concluded with an open discussion with the audience.

Our DPhil student Jesse Wright has been attending the prestigious Dagsthul Seminar Trust and Accountability in Knowledge Graph-Based AI for Self Determination in January 2025.

As a result of this seminar, participants of this seminar co-authored a seminal paper “Towards Computer-Using Personal Agents”.

The paper outlines the vision of Computer-Using Agents (CUA), which are expected to enable users to automate increasingly complex tasks, in the not too far future, using graphical interfaces such as browsers. The authors propose Computer-Using Personal Agents (CUPAs) that have access to an external repository of the user’s personal information, offering users better control of their personal data, the potential to automate more tasks involving personal data, better interoperability with external sources of data, and better capabilities to coordinate with other CUPAs in order to solve collaborative tasks involving the personal data of multiple users.

The full paper can be found arxiv.