Accepted Workshops
Workshops take place on 10–11 May 2027 in Dublin. The conference deadline for workshop paper submissions is 4 January 2027 (AoE). Please check the individual workshop websites for submission instructions.
AccessAgent · AgentDM · AGORA · AiOfAi · AMCH2027 · ATLAS’2027 · BeyondFacts · CodeMates · EARL · FL@FM-TheWebConf’27 · GMLLM’27 · GreenWeb · HCRS@WWW2027 · IDEA · LARS · MM4SG · NSLP-SDP 2027 · PromptEng’27 · SynIRgy · SynWeb · TempWeb · TIME 2027 · TUBLMs · WASST · WebAndTheCity
AccessAgent
AccessAgent: Accessibility for the Agentic Web
Full-day workshop · AccessAgent website
The Web is rapidly evolving into an agent-mediated ecosystem in which intelligent agents increasingly navigate interfaces, retrieve information, complete transactions, and act on behalf of users. While this shift creates transformative opportunities for accessibility through personalized, multimodal, and adaptive assistance, it also introduces new barriers that extend beyond traditional webpage and interface accessibility. In the Agentic Web, accessibility becomes an end-to-end property of the interaction pipeline, encompassing intent understanding, planning, tool use, information retrieval, communication, consent, explanation, action execution, and error recovery. This workshop brings together researchers and practitioners from web accessibility, human-computer interaction, autonomous agents, foundation models, assistive technologies, responsible AI, standards, and disability communities to establish Accessibility for the Agentic Web as an emerging research agenda. The workshop will examine how accessibility can be incorporated into the design, reasoning, personalization, evaluation, and governance of Web agents; identify new accessibility risks and opportunities across diverse user abilities and high-stakes domains; and explore benchmarks, design principles, evaluation methodologies, and standards for accessible agentic systems. Ultimately, the workshop aims to build a multidisciplinary community and develop a shared roadmap toward an Agentic Web that enhances, rather than restricts, user agency, autonomy, and equitable access to digital services.
AgentDM
The First Workshop on Structured Data Mining with LLM Agents
Half-day workshop · AgentDM website
The First Workshop on Structured Data Mining with LLM Agents (AgentDM) brings together researchers and practitioners working on large language model agents, data mining, and structured data management. The workshop explores how agents can plan, use tools, and reason over tables, databases, graphs, time series, and other structured data to support discovery, prediction, and decision-making. Key topics include reliable agent workflows, evidence-grounded reasoning, verification, and evaluation. Through invited talks, paper presentations, and interactive discussions, AgentDM aims to connect the Web, data mining, and agent communities and identify shared challenges and future research directions for trustworthy structured data mining.
AGORA
International Workshop on Architectures, Governance, Orchestration and Reasoning for the Agentic Web
Half-day workshop · AGORA website
The Web is evolving as artificial intelligence becomes not merely a user-oriented technology but an active participant in the Web’s operational processes. Rather than existing as isolated models or applications, modern AI systems are increasingly emerging as connected entities capable of perceiving, reasoning, and acting through standardized Web interfaces. From LLM-based assistants equipped with function calling to API-driven service bots, cognitive copilots, and digital companions, a new generation of Web-integrated intelligence is taking shape. These systems interact with heterogeneous data sources, invoke REST and data APIs, coordinate external tools and services, and update shared resources, transforming the Web into a dynamic environment for contextual reasoning, decision-making, and automation. A key factor behind this transformation is the emergence of open and lightweight protocols that allow AI systems to connect securely with Web services and data sources. Standards and frameworks such as the Model Context Protocol (MCP), OpenAPI specifications, and other API-based connectors illustrate this shift toward standardized and transparent communication between AI agents and external systems. These technologies address long-standing challenges in AI integration by enabling systems to access and use distributed knowledge in a secure and consistent way, without the need to build a custom connector for each platform. By abstracting the complexity of APIs and services, they foster greater interoperability, reliability, and security across the emerging Agentic Web. Treating agents as first-class, governed participants in these Web ecosystems raises a complementary set of questions beyond pure connectivity. Service-oriented computing offers decades of accumulated practice in composing, discovering, and orchestrating autonomous components under explicit contracts and quality guarantees. This body of practice is only beginning to be inherited and adapted by agentic systems to their own needs. At the same time, once agents act with real autonomy over shared infrastructure, questions of trust, accountability, economic incentive, and sustainable long-term operation move from afterthought to design requirement, alongside the need for rigorous ways to evaluate agent reasoning and to keep humans meaningfully in control of agent behavior. This workshop brings these strands together: the technical architectures that let agents operate on the Web, the governance and economic mechanisms that keep that operation trustworthy and accountable, and the reasoning and evaluation practices that let the community understand what agents actually do.
AiOfAi
Workshop on Adverse Impacts and Collateral Effects of Artificial Intelligence Technologies
Half-day workshop · AiOfAi website
Artificial Intelligence (AI) is reshaping the Web not only through the generation of content, but increasingly through systems that can reason, browse, communicate, invoke tools, and act on behalf of users. This transition toward increasingly agentic AI introduces new risks to security, privacy, information integrity, trust, and human autonomy. AI-enabled deception, deepfakes, misinformation, impersonation, and social engineering remain critical concerns, alongside emerging threats such as prompt injection, goal hijacking, unsafe tool use, memory poisoning, and excessive agency. The AiOfAi workshop, building on four previous editions at The International Joint Conference on Artificial Intelligence (IJCAI) and The ACM Web Conference, examines the adverse impacts and collateral effects of AI technologies and the mechanisms needed to understand, evaluate, and mitigate them. The 2027 edition places particular emphasis on the emerging agentic Web, where AI systems increasingly consume information from the Web and autonomously interact with online content, services, tools, and other agents. We seek contributions addressing deceptive and manipulative AI, agentic-AI security, privacy and information leakage, information integrity, identity and delegated authority, trustworthy human–AI interaction, provenance, adversarial evaluation, monitoring, and human oversight.
AiOfAi 2027 will bring together researchers and practitioners from artificial intelligence, Web security and privacy, human–computer interaction, social computing, cybersecurity, and responsible AI. Through invited talks, research presentations, posters and demonstrations, and structured discussions, the workshop will connect technical failure analysis with human and societal impacts and identify mitigation strategies and research priorities for building a safer, more trustworthy, and human-centered AI-powered Web.
AMCH2027
Aging Well with Web and AI Technologies for Maintaining Good Mental and Cognitive Health
Half-day workshop · AMCH2027 website
Populations are ageing faster than care systems can adapt. Older adults increasingly find information, stay connected and receive care through the Web, and now through AI assistants and agents built on it. AMCH 2027 revives the Workshop on Aging Well with Technology for Maintaining Good Mental and Cognitive Health, first held at WWW 2017 in Perth. We dedicate it to the memory of its co-founder, Professor Mark Chignell (1956–2025). The half-day, in-person workshop brings together Web, AI, HCI and health re- searchers to (i) share new results on Web- and AI-based assessment and promotion of mental and cognitive health, (ii) address the trust, safety, privacy and inclusion problems raised by AI companions and agents for older adults, and (iii) draft a shared research roadmap in structured interactive sessions.
ATLAS’2027
Second International Workshop on Agents, Things and Lightweight AI for Sustainability
Full-day workshop · ATLAS’2027 website
With the strong growth of IoT applications and the mainstreaming of agentic AI, the environmental and societal impact of billions of connected devices, as well as the increasing use of energy-intensive algorithms demand urgent attention. The second edition of the ATLAS workshop aims to offer a platform for scholars and practitioners to exchange experiences and present ongoing work on how resource limitation, sustainability and frugality should be taken into account while designing, managing and optimizing agentic AI-based and physical infrastructures. It also aims to review existing implementations of the Web of Things (WoT) standards with respect to long-term maintainability and environmental awareness.
BeyondFacts
7th International Workshop on Computational Methods for Online Discourse Analysis
Full-day workshop
Online discourse offers valuable insights into misinformation, bias, and societal debates, yet its computational analysis requires shared definitions of concepts such as claims, arguments, stances, and veracity. This workshop brings together communication studies, computational linguistics, and computer science to advance the modeling, extraction, and analysis of discourse on the Web. It explores approaches combining symbolic knowledge representation with statistical AI, including machine learning and large language models. By promoting interdisciplinary collaboration and reusable, structured discourse data, the workshop supports scientific and journalistic investigations of online debates.
CodeMates
Next-Gen Code Development with Collaborative AI Agents
Full-day workshop
The second edition of Next-Gen Code Development with Collaborative AI Agents (CodeMates) explores how coding agents collaborate with human developers and with one another to build reliable software. Bringing together researchers and practitioners from different disciplines including AI, software engineering, human-computer interaction, and Web systems, the workshop focuses on agent coordination, human-AI collaboration, and the evaluation and deployment of trustworthy coding agents at Web scale.
EARL
EARL: The 4th Workshop on Evaluating and Applying Recommender Systems with Large Language Models
Half-day workshop · EARL website
The 4th Workshop on Evaluating and Applying Recommender Systems with Large Language Models (EARL), co-located with The ACM Web Conference 2027 in Dublin, will bring together researchers and industry practitioners working at the intersection of large language models (LLMs) and recommender systems. Building on the success of previous editions, EARL 2027 will focus on how to effectively evaluate, apply, and deploy LLM-based recommendation approaches, with particular attention to real-world challenges, reproducibility, and responsible use. The workshop aims to encourage discussions, share practical insights, and explore future research directions in this rapidly evolving field.
FL@FM-TheWebConf’27
International Workshop on Federated Foundation Models for the Web 2027
Half-day workshop
The rise of foundation models (FMs) amplifies the importance and relevance of federated learning (FL) as a crucial research direction. With FMs becoming the norm in machine learning development, the focus shifts from model architecture design to tackling the issues surrounding privacy-preserving and distributed learning. Advancements in FL methods have the potential to unlock the use of FMs, enabling efficient and scalable training while safeguarding sensitive data. With this in mind, we invite original research contributions, position papers, and work-in-progress reports on various aspects of federated learning in the era of foundation models. Since the emergence of foundation models has been a relatively recent phenomenon, their full impact on federated learning has not yet been well explored or understood. We hope to provide a platform to facilitate interaction among students, scholars, and industry professionals from around the world to discuss the latest advancements, share insights, and identify future directions in this exciting field.
GMLLM’27
The 4th Frontiers in Graph Machine Learning for the Large Model Era
Half-day workshop · GMLLM’27 website
The “4-th Frontiers in Graph Machine Learning for the Large Model Era” workshop focuses on advancing graph machine learning (GML) techniques in the context of increasingly large and powerful models. Graphs offer a principled way to represent structured and relational data, making them essential for capturing complex dependencies in knowledge, systems, and behaviors. As the scale and influence of foundation models grow, graph learning stands at a unique vantage point to enhance model robustness, improve interpretability, and integrate domain-specific relational priors. This workshop explores how graph learning can support emerging needs in knowledge reasoning, temporal and multi-hop inference, and AI systems. It also investigates how advances in representation learning, structure-aware generalization, and efficient graph processing can contribute to trustworthy and scalable AI systems. By convening experts in graph learning, knowledge management, and LLMs, the workshop aims to identify core challenges and opportunities of GML in the large model era.
GreenWeb
GreenWeb 2027: First International Workshop on Sustainable and Energy-Efficient Web and AI
Half-day workshop · GreenWeb website
The web now carries a measurable share of the world’s electricity demand, and the arrival of large language models in everyday web products is pushing that share up quickly. GreenWeb 2027 is a half-day workshop on measuring and reducing the energy and carbon footprint of websites, web systems, and the AI services behind them. It brings together researchers, practitioners, and standards contributors, and combines peer-reviewed papers and posters with breakout groups that produce a community research roadmap and an open catalogue of measurement tools and datasets. It is the first workshop at The Web Conference on the sustainability of the web itself.
HCRS@WWW2027
The 3rd Workshop on Human-Centered Recommender Systems
Half-day workshop · HCRS@WWW2027 website
Recommender systems now shape what people know, choose, and experience. In the generative AI era, recommendation is becoming an ongoing relationship among people, agents, and platforms. User-facing agents can interpret goals, maintain memory, use tools, and act across multiple steps, while agentic workflows can revise, evaluate, and deploy recommendation models. These capabilities promise more adaptive support, but they may also shift control over how preferences are represented, decisions are executed, and systems evolve. Human-centered recommendation must therefore ask whose goals agents serve, how people can inspect and correct their behavior, and whether their effects remain beneficial over time. Organized around agency, alignment, and long-term value, the 3rd Workshop on Human-Centered Recommender Systems (HCRS@WWW 2027) brings together recommender systems, information retrieval, human-computer interaction, responsible AI, and social computing to develop methods, evaluations, and safeguards that keep agent-mediated recommendation accountable to people.
IDEA
The First International Workshop on Item IDentifiers Design, Evaluation, and Adaptation
Half-day workshop · IDEA website
Web search, recommendation, and advertising connect users to changing collections of documents, products, and media. An item ID can be a lookup key, a representation shared across similar items, or a sequence that determines which candidates a model can retrieve. These roles raise a concrete design question: when should an ID share information, preserve item-specific distinctions, or remain stable as a Web catalog changes? Semantic IDs sharpen this question as they spread across generative retrieval, ranking, joint search and recommendation, and advertising. Yet evidence from one setting may not establish whether an ID design will work in another. Better retrieval or ranking performance may also reflect changes to input features, downstream models, or evaluation protocols. As catalogs evolve, new items must be incorporated without destabilizing existing representations. We propose , the First International Workshop on Item Identifiers Design, Evaluation, and Adaptation. This half-day workshop brings together designers of raw, semantic, and hybrid IDs, evaluation researchers, and systems engineers around three questions: what should IDs encode, what evidence justifies an ID choice, and what must survive when representations are updated? Negative results, controlled comparisons, and lessons from catalog updates are seldom gathered in one venue. IDEA brings them together: Participants would examine competing positions against shared cases and produce a public Community Roadmap, summarizing findings, priorities for follow-up studies, and collaboration resources.
LARS
The Second Workshop on LLM and Agents for Recommendation Systems
Full-day workshop
Large language models (LLMs) and AI agents are reshaping recommendation systems, extending their role beyond semantic understanding and content generation toward generative recommendation, reasoning, personalization, memory, and autonomous decision making. Following the successful first edition of LARS at The Web Conference 2026, the Second Workshop on LLM and Agents for Recommendation Systems (LARS 2027) brings together researchers and practitioners to study this rapidly evolving intersection. The workshop broadens its scope to cover both LLM-inspired and LLM-native recommendation, agentic systems with retrieval-augmented generation (RAG), memory and tool use, multimodal and conversational recommendation, and recommendation beyond next-item prediction, including slate and whole-page optimization and long-horizon user experiences. LARS also emphasizes rigorous evaluation, efficiency, responsible recommendation, and the challenges of deploying these approaches in real-world Web-scale systems. Through invited talks, contributed papers, interactive discussions, and breakout sessions, LARS aims to connect advances in foundation models and agents with fundamental recommendation problems and foster new research directions, benchmarks, and practical applications for the next generation of recommender systems.
MM4SG
The Fifth Workshop on Multimodal Content Analysis for Social Good (MM4SG 2027)
Half-day workshop · MM4SG website
The Fifth Workshop on Multimodal Content Analysis for Social Good (MM4SG 2027) will bring together researchers and practitioners to explore emerging challenges and innovative solutions for analysing and moderating multimodal content across digital platforms. The workshop will focus on understanding text, images, memes, and other multimodal content, addressing issues such as misinformation, harmful content, authenticity, and the growing impact of AI-generated and fabricated media enabled by Large Language Models (LLMs). By fostering interdisciplinary discussions across natural language processing, machine learning, computational social science, and ethics, MM4SG 2027 aims to advance reliable, responsible, and effective multimodal content analysis for social good.
NSLP-SDP 2027
Workshop on AI for the Scholarly Web
Full-day workshop · NSLP-SDP 2027 website
The scholarly Web has become the primary infrastructure through which scientific knowledge is published, discovered, linked, assessed, and reused. At the same time, advances in large language models, multimodal foundation models, knowledge graphs, retrieval-augmented generation, and agentic AI are creating new opportunities for transforming scholarly content into machine-actionable knowledge and for building AI systems that can search, interpret, synthesize, and verify scientific information. NSLP-SDP 2027 – Workshop on AI for the Scholarly Web – brings together the previously separate Natural Scientific Language Processing (NSLP) and Scholarly Document Processing (SDP) communities to address this full scholarly-information pipeline. The workshop spans scientific document understanding, information extraction and linking, scholarly and research knowledge graphs, Web-scale search and retrieval, multimodal and multilingual processing, scientific question answering and synthesis, provenance and evidence grounding, and trustworthy research agents. A central emphasis is placed on robust evaluation, reusable resources, and four shared tasks covering cross-lingual scientific fact-checking, scholarly entity and relation extraction, research-artifact recognition in software repositories, multimodal scientific question answering, and software-mention extraction and coreference. By connecting researchers from Web science, NLP, information retrieval, digital libraries, knowledge representation, scientometrics, scholarly communication, and research infrastructures, NSLP-SDP 2027 aims to establish a cross-disciplinary forum for turning the fragmented scholarly Web into a trustworthy, interoperable, and machine-actionable knowledge environment for humans and AI agents.
PromptEng’27
Fourth International Workshop on Prompt Engineering for Pre-Trained Language Models
Half-day workshop · PromptEng’27 website
The recent achievements and availability of Large Language Models have paved the road to a new range of applications and use-cases. Pre-trained language models are now being involved at-scale across numerous fields, where they were previously unutilized. More specifically, the recent progress in generative models has paved the way to using them easily through textual instructions aka. prompts. Unfortunately, the performances of these models are highly dependent on the exact phrasing used in prompts and therefore practitioners need to adopt fail-retry strategies.
This fourth international workshop on prompt engineering aims at gathering practitioners (both from Academia and Industry) to exchange about good practices, optimizations, results and novel paradigms about the designing of efficient prompts and context-building to make use of LLMs, with a specific focus this year on how they impact agent performances.
SynIRgy
SynIRgy: Simulation and Synthetic Data for Web-Scale Information Access
Half-day workshop · SynIRgy website
Evaluating today’s information access systems is hard when it relies only on logged user data, which tends to be skewed, incomplete, and often restricted for privacy reasons. Simulation and synthetic data offer a way forward: they create controlled, repeatable settings in which web search, recommendation, and personalization algorithms can be trained, tested, and compared. SynIRgy at TheWebConf 2027 invites researchers and practitioners to exchange experience, compare practices, and work on open problems in the use of simulation and synthetic data for information access on the Web. It follows the first SynIRgy workshop at ECIR 2026 and brings the topic to the web community. We welcome methodological work, ethical reflection, and applied case studies. The program is built around discussion rather than paper presentations alone, and it will lead to a short report and a community roadmap for the field.
SynWeb
SynWeb: 1st Workshop on Trustworthy Synthetic Data for the Web: Generation, Verification, and Governance
Half-day workshop · SynWeb website
SynWeb is a half-day workshop on trustworthy synthetic data for the web. Large language and multimodal models are increasingly used to create training corpora, benchmarks, user and item profiles, structured records, and evaluation data for web-facing systems. At the same time, machine-generated content is becoming part of the web itself, raising questions about provenance, contamination, distribution shift, privacy, and whether synthetic data actually improves downstream web tasks. SynWeb brings together researchers and practitioners in web data management, information retrieval, NLP, recommender systems, and responsible AI to study this problem end to end: how synthetic web data should be generated, verified, documented, stress-tested, and evaluated before it is reused. The workshop will combine technical contributions with a benchmark-design session and a community roadmap aimed at producing concrete evaluation criteria and shared research directions for trustworthy synthetic data on the web.
TempWeb
The 17th TemporalWeb Analytics Workshop
Full-day workshop · TempWeb website
TempWeb focuses on investigating infrastructures, scalable methods, and innovative software for aggregating, querying, and analyzing heterogeneous data at Internet scale. Emphasis will be given to temporal data analysis along the time dimension for web data that has been collected over extended time periods. The objective of this workshop is to provide a venue for researchers of all domains (IR, NLP, Web mining, etc.) where the temporal dimension opens an entirely new range of challenges and possibilities. The workshop’s ambition is to keep shaping a community of interest on the research challenges and possibilities resulting from the introduction of the time dimension in web analysis and related topics.
TIME 2027
TIME 2027: 3rd International Workshop on Transformative Insights in Multi-faceted Evaluation
Full-day workshop · TIME 2027 website
Welcome to TIME 2027 (3rd International Workshop on Transformative Insights in Multi-faceted Evaluation), held in conjunction with The Web Conference 2027. As foundation models, multimodal AI, generative systems, retrieval-augmented generation (RAG), and AI agents become deeply embedded across the Web, powering search, recommendation, content creation, and autonomous workflows, their capabilities and risks can no longer be captured by conventional task-specific accuracy metrics alone. Operating across heterogeneous data, external tools, and dynamic user environments introduces critical challenges such as hallucination, bias, privacy leakage, and poor calibration. Building on the successful editions of TIME 2025 and TIME 2026, this full-day workshop brings together researchers and practitioners to address a central question for the next generation of Web AI: how should we evaluate systems whose capabilities, behaviours, and risks extend far beyond conventional task-level performance? Through invited talks, paper presentations, posters, an interactive panel, and breakout discussions, TIME 2027 provides a collaborative forum to examine what should be evaluated, how it should be measured, and how evaluation insights translate into trustworthy, real-world Web systems.
TUBLMs
Trust Under Budget: Reliable, Efficient, and Safe (Vision) Large Language Models for Deployment
Full-day workshop · TUBLMs website
Deploying (vision-)large language models in real-world Web systems demands more than benchmark accuracy: answers must stay faithful, agents must resist manipulation, and safety checks must fit within serving budgets. In production, reliability, efficiency, and safety are not independent goals—they are jointly bounded by explicit token, compute, memory, latency, energy, and monetary budgets. This full-day workshop introduces a unifying "trust under budget" framework organized around three pillars: (1) reliable reasoning under constrained tokens and compute; (2) efficient inference that meets latency, memory, and cost limits; and (3) safe deployment that preserves alignment after quantization, distillation, and compression. By connecting model-level methods to service outcomes such as answer quality, agent reliability, moderation robustness, and serving cost, the workshop bridges the AI research and deployment communities and advances benchmarks and evaluation protocols for (v)LLMs that are both capable and certifiably trustworthy under industrial constraints.
WASST
WASST 2027: The 1st Workshop on Security, Safety, and Trust for the Agents on the Web
Half-day workshop
A growing share of web activity is now performed by AI agents acting on people’s behalf. This is reshaping the risk landscape of the web. Agents can make mistakes with real-world consequences, be manipulated through untrusted web content, or be deliberately used to defraud users and abuse web services. Addressing these risks requires expertise across multiple research disciplines. WASST 2027 brings together researchers and practitioners from academia and industry across these disciplines, focusing on three themes. Safety covers agent failures that occur without an adversary. Security covers attacks that hijack or manipulate agents through web content, tools and interfaces. Trust covers malicious agents and how to distinguish them from legitimate ones. We welcome contributions on methods, evaluation, resources, governance and societal implications. We particularly encourage contributions from early-career researchers and industry.
WebAndTheCity
13th WebAndTheCity: the Web for AI-Driven Citiverse
Half-day workshop · WebAndTheCity website
This is the 13th edition of the annual workshop series labeled “WebAndTheCity – The Web and Smart Cities”, which is the successor of the series of workshops that started in Florence in 2015 and has continued taking place every year in conjunction with The Web Conference series. Each year, the focus of the workshop is actualized, and this year, the workshop focuses on AI-driven Citiverse. In the era of AI and the citiverse (metaverse for people-centric cities), cities are being transformed into intelligent urban environments where utility and public services are being transformed to autonomous, fed by multiple data sources and supported by analytics and agents. This workshop aims to explore how the Web supports this transformation and how its technologies can improve the smart cities and their integrated digital and physical worlds.
