The Web is changing faster than at any time since its creation. Generative AI now produces a growing share of Web content, autonomous agents browse, negotiate, and transact on behalf of people, and long-held assumptions about who reads, writes, and pays for the Web are being reopened. This track invites papers that look further ahead than a regular research paper can. We seek bold, forward-looking, and possibly controversial ideas that could change how the Web is built, used, governed, or studied over the next decade.

A Brave New Ideas paper should make a concrete research contribution around an idea that is genuinely new, unconventional, or forward-looking. It may introduce a new problem formulation, technical approach, system architecture, theoretical framework, methodology, or other research contribution that could open a substantially new direction for Web research. It may also challenge a widely held assumption or bring concepts and methods from other disciplines to bear on Web research in fundamentally new ways.

Because such ideas may be at an early stage, we do not expect the same level of experimental maturity or comprehensive evaluation as in the regular research tracks. However, submissions should provide sufficient technical, theoretical, or empirical substance to demonstrate that the proposed idea is more than a vision. This may take the form of preliminary experiments, a prototype or proof of concept, theoretical or formal analysis, a new dataset or methodology, or other appropriate evidence establishing the plausibility and research potential of the idea. Submissions will be judged on the originality and significance of the research contribution, the soundness of the underlying reasoning, the strength of the supporting evidence, and the potential to open important new directions for the Web research community.

This track is not a home for incremental improvements, surveys, incomplete regular research papers, or vision and perspective papers whose primary contribution is an argument or research agenda. A bold thesis alone is not sufficient: submissions must contain an identifiable research contribution and provide appropriate evidence or analysis to substantiate it. Authors must state on the first page what is brave and new about their research contribution and why it matters for the future of the Web.

We expect submissions to present genuinely new ideas grounded in the insightful vision and intellectual contributions of human researchers, rather than relying on AI-generated or otherwise unsubstantiated content. The Chairs and Review Committee will carefully assess the quality of the writing, the clarity of the arguments, and the soundness of the underlying justification and logic. We place strong value on clear, human-understandable communication, as well as the meaningful accumulation and dissemination of knowledge through scholarly papers.

The relevant topics include, but are not limited to:

• Visions and architectures for a Web shared by people and AI agents
• Rethinking core Web concepts such as pages, links, browsers, search, and open standards when content is increasingly generated and consumed by machines
• New economic and incentive models for creators, publishers, and platforms when traffic comes from AI systems and agents as well as people
• New foundations for trust, identity, provenance, and authenticity on a Web with abundant synthetic content
• Alternative futures for a decentralized, user-owned, or community-governed Web
• New ways for people to find information and act on the Web beyond queries, pages, and feeds
• The Web as a substrate for collective intelligence, democratic discourse, and scientific discovery
• Long-term societal risks and opportunities of the Web, including the sustainability of open knowledge commons
• Critical reflections that challenge established methods, evaluation practices, or assumptions in Web research
• Cross-disciplinary ideas that bring law, economics, the social sciences, the humanities, or design to bear on the future of the Web
• Grand challenges and research agendas for the Web community

The Web is a platform for economic activity at a very large scale, from online marketplaces and advertising to cloud services and the sharing and creator economies, and it increasingly shapes how societies work, trade, and organize. Generative AI and autonomous agents are now changing these markets: AI systems stand between consumers and sellers, and agents are starting to search, bargain, and buy on behalf of people and organizations. This track is a forum for theoretical, empirical, and applied research on the modeling, analysis, and design of Web-related economic activity and online markets, and on their wider effects on digital society.

The relevant topics include, but are not limited to:

• Advertising auctions, pricing, markets, and exchanges
• Mechanism and market design for online platforms and marketplaces
• Economics and fairness of platforms, search, and recommender systems
• E-commerce, the sharing economy, the gig economy, and the creator economy
• Economics of generative AI infrastructure and applications, including cost models for LLMs in production systems
• Uses of LLMs and generative AI for marketplace design, bidding, and strategic interaction
• Agent-mediated commerce: markets in which AI agents search, negotiate, and transact on behalf of people and organizations
• Incentives, strategic behavior, and equilibria in multi-agent and human-AI economic environments
• Economics of Web content and data, including creator compensation, data markets, licensing, and the value of training data
• Economic aspects and information design of the Web, generative AI, and cloud computing
• Incentives in network design for Web infrastructures and ecosystems
• Social networks and social learning in economic settings
• Economic aspects of blockchains, cryptocurrencies, and decentralized finance
• Digital labor, online work, and the effects of the Web and AI on employment
• Digital inequality, access, and inclusion in the online economy
• Fairness, privacy, and diversity in economic environments
• Economic analysis of competition, regulation, and policy for digital platforms and markets
• Sustainability of Web economics

Graphs provide a natural formalism for a substantial portion of the Web, encompassing hyperlink structures, social and communication networks, user–item interactions, and knowledge graphs, as well as the interaction networks that have more recently emerged among AI agents, tools, and services. This track solicits original research addressing the scientific challenges associated with graphs on the Web, including but not limited to graph models and algorithms, graph mining and analysis, representation learning, and practical applications.

The relevant topics include, but are not limited to:

• Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Querying, indexing, and retrieval in Web-related graphs
• Efficient manipulation of static and dynamic Web-related graphs
• Scalable, parallel, distributed, and streaming algorithms for Web-scale graphs
• Random graph models and generative models for Web-related graphs
• Graph embeddings and representation learning for Web-related graphs
• Graph neural networks and graph neural architecture search approaches for Web-related graphs
• Robust, explainable, and privacy-preserving learning on Web-related graphs
• Graph foundation models and graph LLMs for Web-related graphs, including graph-based information processing and intelligence
• Modeling and analysis of graphs that arise from AI agents on the Web, such as agent-tool, agent-agent, and human-agent interaction networks

Over the past decades, the Web and the internet have grown to be an integral part of our lives. As generative AI and autonomous agents take on more of what happens on the Web, questions of accountability, consent, and human agency become more pressing. Through this track, we aim to surface the latest research on what it means to build a responsible Web and internet, one that contributes towards more equitable, emancipatory, and sustainable futures.

The relevant topics include, but are not limited to:

• Algorithmic accountability and transparency on the Web
• Audits and impact assessments of Web technologies, standards, platforms, and applications, including systemic risks and access to evidence for independent scrutiny
• Accountability, oversight, and auditing of AI agents that act on the Web on behalf of people and organizations
• Consent frameworks and practices on the Web
• Privacy and data protection on the Web, including privacy-enhancing technologies and the consequences of tracking and profiling
• Ethical and legal aspects of Web-scale data collection, analysis, and use, including the use of Web content to train and ground AI systems
• Fairness, accountability, transparency, and ethics of Web technologies
• Individual and societal consequences of algorithmic deployment on the Web, including lived experiences, distributional effects, and long-term impacts
• Human agency and meaningful control over Web systems, including contestability, recourse, and redress, as well as control over tasks delegated to agents
• Measurement, analysis, and circumvention of Web censorship
• Information integrity on the Web, including misinformation and disinformation, content provenance, and the effects of AI-generated content
• Mitigating online harms, including harassment, spam, and fake reviews
• Deceptive and manipulative design on the Web, digital well-being, and age-appropriate services that respect children’s rights
• Social and technical mechanisms of refusal for Web technologies and applications, including the means for creators and site owners to control AI access to their content
• Accessibility and inclusion on the Web
• Sustainability and climate impact of Web technologies
• Governance, policy and regulation of Web technologies, including AI-agent systems acting on the Web
• Content moderation and platform governance, including enforcement practices, appeals, and their consequences for affected communities
• Web for democratic and emancipatory futures

This track focuses on the foundations and applications of search, recommendation, and retrieval-augmented AI on the Web. We invite original, high-quality contributions on models, algorithms, and systems for search and ranking, query understanding, recommendation, interactive and conversational information access, and multilingual and domain-specific search. The track also covers the directions in which information access is now moving: ranking and recommendation with LLMs, retrieval-augmented generation (RAG), multi-modal RAG, and agentic search, in which AI systems plan, issue queries, and synthesize results over many steps. Our goal is to foster research that connects established information retrieval and recommender systems research with the capabilities of modern AI.

The relevant topics include, but are not limited to:

• Web query analysis, representation, and understanding
• Web search models and ranking
• Web crawling and indexing
• Efficiency and scalability of Web search engines and recommender systems
• Learning to rank, online learning, and counterfactual learning for ranking and recommendation
• Recommendation models and algorithms, including collaborative, content-based, sequential, session-based, and cross-domain recommendation
• Generative retrieval and generative recommendation
• Large language models for search and recommendation
• Assisted, interactive, and conversational search and recommendation
• Context-aware and cross-device search and recommendation
• Vertical and domain-specific search and recommendation
• Multilingual and cross-lingual Web search
• Web question answering
• Natural language understanding for Web search
• Ad search, and search and recommendation for Web retail
• Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Agentic search and deep research systems that plan and carry out multi-step information tasks
• Retrieval and ranking for AI agents as consumers of results, including retrieval of tools, services, and long-term memory
• Evaluation of search, recommendation, and retrieval-augmented systems

This track offers researchers the opportunity to present their work to the broad community interested in security and privacy on the Web. We invite submissions that address any aspect of Web security and privacy, including theoretical foundations, practical implementations, measurements, and case studies. We also welcome work on the new attack surface created when AI agents read untrusted Web content and act with their users’ authority.

The relevant topics include, but are not limited to:

• Applications of cryptography
• Authentication, authorization, and access control
• Blockchains and distributed ledgers
• Browser and end-point security
• Cyber-crime defenses and forensics
• Cryptocurrency and smart contracts
• Data security and privacy
• Data transparency and provenance
• Large-scale security measurements
• Security in Web-based applications and services
• Tracking, profiling, and countermeasures against them
• Privacy-enhancing technologies
• Security and privacy of machine learning and AI applications
• Security of AI agents on the Web, including prompt injection through Web content, tool misuse, and isolation of browsing and computer-use agents
• Identity, authentication, and delegated authority for agents acting on behalf of users
• Detection and management of automated traffic, including bots, scrapers, and AI agents
• AI-enabled attacks and defenses, such as phishing, fraud, and social engineering at scale
• Privacy of personal data that users share with AI assistants and agents

This track is a forum for researchers interested in knowledge graphs and other structured data models with machine-interpretable semantics, which are widely adopted in advanced applications on the Web. We invite original research on methods, algorithms, techniques, and applications that support the creation, acquisition, publication, and consumption of interlinked structured data on the Web, the (human-assisted) semantic integration of data, and the enrichment and processing of large, real-world datasets in a Web context. Machine-readable semantics were conceived with software agents in mind, and the arrival of LLM-based agents gives new weight to explicit, shared, and verifiable knowledge that both people and machines can rely on.

The relevant topics include, but are not limited to:

• Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable knowledge graphs and ontologies
• Representation, semantic annotation, enrichment, access, and integration of a variety of data on the Web, including semi-structured data, tabular data, text, multimedia (images, sound, video), and sensor data
• Methods, algorithms, and applications for the development of semantic models, knowledge graphs, and other structured data models with machine-interpretable semantics
• Methods that combine semantic models such as knowledge graphs with LLMs to extend what either can do, including knowledge-grounded generation and neuro-symbolic reasoning
• Data modeling to support human-machine intelligence, including LLM agents, intelligent system behavior, explanations, and user-friendly interaction
• Knowledge graphs and structured knowledge as grounding, memory, and a means of verification for LLMs and agents
• Machine-readable descriptions of Web content, services, APIs, and tools that allow agents to discover, understand, and use them
• Knowledge extraction, entity linking, and knowledge graph construction and completion with LLMs
• Provenance, trust, security and privacy, and ethical issues in managing semantic data
• Applications of semantic technologies for improving search, browsing, recommendation, and personalization in domains of interest to the Web community, such as news, misinformation detection, open science, medical and health science, the humanities, and the social sciences

We welcome submissions in all areas that concern social networks, social media, and the interaction of the Web and society. This includes online social spaces in which AI-generated content and AI agents take part alongside people.

The relevant topics include, but are not limited to:

• Social media analysis through the lens of networks
• Social mining and social search on the Web
• Influence propagation, information diffusion, and prediction on networks (link, node, or subgraph prediction)
• Socio-technical Web-based systems and communities
• Opinion dynamics on the Web and social media (e.g., filter bubbles, polarization, conflicts, or tribalization)
• Media and governance (self-governance, civic engagement, and collective action via the Web)
• Analysis of Web-mediated communications, communities, and crowds
• Coordinated and inauthentic behavior, social bots, and manipulation campaigns
• Fairness and bias in social network and social media analysis
• Generative AI and large language models and their impact on social systems
• AI agents as participants in social networks and online communities, and the interactions between people and agents
• Simulation of social systems and collective behavior with LLM-based agents, and the validity of such simulations
• Decentralized and federated social platforms
• Applications and emergent phenomena

This track invites original research on the design, construction, and performance evaluation of Web-based, mobile, and ubiquitous computing systems and architectures, including Web of Things (WoT) systems. It also covers the systems that allow AI agents to operate on the Web: the platforms that run them, the protocols and interfaces through which they reach sites, services, and each other, and the infrastructure that serves them. We welcome submissions of interest to both academic and industrial research that advance the understanding of these systems, for example through data analysis, empirical studies, novel application designs and implementations, deployment experiences and lessons learned, and usability evaluations.

The relevant topics include, but are not limited to:

• Cloud, edge, and content delivery systems for the Web
• Decentralized Web and Fediverse systems
• Web performance, measurement, and characterization
• Virtualization and resource management in Web systems and infrastructures
• ML and AI systems for Web, mobile, and WoT, including applied ML and AI for Web-based mobile applications
• Federated Web and WoT systems, including distributed, federated, and edge-based data processing
• Location- and context-aware Web and WoT applications and services
• Energy management for devices in mobile Web and WoT environments
• Novel mobile and WoT systems, and systems of systems
• Data management and stream processing for Web, mobile, and wireless applications
• Web applications in cross-disciplinary domains and verticals such as mixed reality, smart cities, and digital health
• Browsers, Web engines, and Web standards, including their evolution to support automated and agent-driven use
• Architectures, runtimes, and platforms for building, deploying, and orchestrating AI agents and multi-agent systems on the Web
• Protocols, interfaces, and standards through which agents connect to tools, Web services, and other agents
• Agent-ready Web sites, APIs, and services
• Serving infrastructure for LLMs and agents across cloud, edge, and device, including efficient inference and caching
• Performance, reliability, cost, and observability of agentic systems in production
• Effects of crawler and agent traffic on Web infrastructure, and techniques to manage it
• Experiences and lessons learned from deployments of Web-based algorithms and systems
• Sustainability and carbon-aware systems for Web, mobile, and WoT

In this track, we invite original research on all aspects of inclusive, human-centered approaches to user modeling and personalization on the Web (note this important distinction from other tracks that include agentic systems). The users of the Web are also changing: people increasingly reach the Web through AI assistants and delegate tasks to agents, and those agents become users of Web sites and services in their own right. We therefore welcome work on understanding and modeling people who use the Web through agents, and on agents that represent, simulate, or act for users. Contributions may address data analytics, algorithm development, system design, user studies, use case reviews, and evaluation within the context of the Web.

The relevant topics include, but are not limited to:

• User model development and evaluation
• Studies of user behavior, including longitudinal effects of personalized systems
• Practical large-scale studies of user experience
• Psychology-informed user models and personalization
• Personalized content ranking and presentation
• Personalization for the social Web
• Metrics for user behavior and for evaluating success
• Explainable and interpretable methods for personalization
• Fairness, accountability, transparency, and ethics for personalization
• User privacy protection in personalized systems
• Federated and on-device user modeling and personalization
• User modeling for targeted and personalized online advertising
• Large language models for user modeling and personalization
• User modeling and simulation for interactive and conversational systems, including LLM-based user simulators and their fidelity
• Personal AI assistants and agents that act on behalf of users: preference elicitation, long-term memory, and user profiles for agents
• Delegation, trust, and control between users and their agents, including how users express intent and correct agent behavior
• Studies of how people use the Web through AI assistants and agents, and how this changes browsing, search, and decision making
• Characterizing and modeling the behavior of AI agents as Web users, and telling it apart from human behavior in interaction data
• Personalization and interface design for Web sites and services visited by both people and agents

This track welcomes original, high-quality research on the extraction of information from the Web and on the acquisition, curation, analysis and mining of Web content. Web content is multimodal and multilingual: images, video, audio, and text in thousands of languages and many cultural contexts. A growing share of it is produced by generative models and accessed by AI agents. We invite work that analyzes this content in all its forms, including the methods that let both people and agents make sense of it.

The relevant topics include, but are not limited to:

• Web data integration and cleaning
• Web data generation and simulation
• Web data visualization
• Large pretrained models with Web data
• Machine learning and data science for the Web
• Robustness and generalizability of Web mining methods
• Normalization, clustering, classification, and summarization of Web text
• Topic discovery and tracking
• Community question answering
• Sentiment analysis and opinion mining
• Content-based information diffusion
• Mining and understanding of multimedia Web content, including images, video, audio, memes, short-form video, and live streams
• Multimodal learning and vision-language models for Web content
• Multilingual and cross-lingual analysis of Web content, including low-resource languages, code-switching, and cultural context
• Bridging structured and unstructured data
• Web page understanding and content extraction from markup, rendered layout, and screenshots, including for use by AI agents
• Agentic methods and harnesses for Web content acquisition, extraction, and analysis, including tool use, multimodal context management, and evidence grounding
• Web traffic and log analysis, including the mining of agent interaction traces
• Web measurements
• Models for Web evolution
• Interdisciplinary scientific discovery through Web data mining
• Web data provenance, reliability, and authenticity
• Detection, attribution, and measurement of AI-generated text, images, audio, and video on the Web
• Web data quality in the era of algorithmically generated content

Progress in Web research depends on sound evaluation, on human judgment collected at scale, and on shared datasets, benchmarks, and tools. All three are under pressure. LLMs now act as annotators and judges, benchmarks saturate quickly or leak into training data, and Web agents need test environments that are realistic, safe, and repeatable. This track is a forum for research on how Web systems and Web research should be evaluated, on human and hybrid human-AI computation, and on resources that the community can build on.

We invite three kinds of contribution: methodological work on evaluation and measurement; research on crowdsourcing and human computation; and resource papers that present new datasets, benchmarks, test collections, simulators, or software. Resource papers are assessed on the usefulness of the resource to the Web community, its quality and novelty, its documentation, and its availability; the resource should be accessible to reviewers at submission time.

The relevant topics include, but are not limited to:

• Evaluation methodologies, metrics, and experimental design for Web systems and applications
• Online evaluation, including A/B testing, interleaving, and counterfactual and off-policy evaluation
• Evaluation of generative and LLM-based Web systems, including factuality, attribution, and usefulness
• LLMs as judges and assessors: validity, reliability, and bias relative to human judgment
• Benchmarks, environments, and simulators for evaluating Web agents, covering task success, safety, and cost
• User simulation and synthetic data for evaluation
• Meta-evaluation, statistical validity, benchmark saturation, and data contamination
• Reproducibility, replicability, and generalizability studies of Web research
• Crowdsourcing and human computation: task design, workflows, incentives, and quality control
• Humans versus LLMs for data annotation and labeling (e.g., quality, trade-offs, biases, evaluation)
• Architectures and workflows that use LLMs in crowd work (e.g., productivity, efficiency, quality, biases, oversight)
• Trust and reliance of crowd workers and data experts on generative AI
• Fairness and ethical considerations in crowd work and in human-in-the-loop AI systems
• Data quality of human-annotated and AI-annotated datasets, including annotator disagreement and bias
• New datasets, test collections, and corpora of Web data
• Open-source software, tools, and platforms for Web research
• Documentation, licensing, maintenance, and ethical collection of Web research data