We are again thrilled to be able to present the Computer-Supported Cooperative Work and Social Computing (CSCW) community with an issue of the Proceedings of the ACM on Human-Computer Interaction, containing very interesting and relevant scholarship from its members. This issue includes 42 papers from the May 2025 cycle, selected from a total of 637 submissions and following two rounds of reviews and one revision. 209 submissions from this round will be further revised, reviewed again, and may appear in another issue of the journal later this year. Our external reviewers and track editorial board have together conducted a rigorous review process to select contributions of the highest quality advancing the CSCW field. As Track Chairs, we are grateful for the community’s collective efforts to continue shaping and sharing CSCW’s tradition of high-quality scholarship across the years.

Large Language Models (LLMs) are widely used today in virtual assistants and content generation. However, there are suspicions that LLMs present confirmation bias, responding in a way that reinforces beliefs or assumptions embedded in users' questions, which can lead to erroneous decision-making, especially in sensitive areas such as healthcare. The objective of this research is to determine how often and under what conditions LLMs present confirmation bias and to identify the causes of this effect. The methodology involves conducting an experiment in which 52 biased healthcare questions are presented to 10 of the most popular models and analyzing whether their responses were biased. This work proves with statistical power the behavior of confirmation bias. We show that confirmation bias in LLMs occurs in all LLMs with a frequency of 20% to 60% of the occasions. The evidence suggests that the bias arises from the training database, the Transformer architecture itself, and the instructions in the fine-tuning phase by the companies behind the LLMs. This research explores pathways for the development of trustworthy LLMs.

Software development teams depend on the constant and varied use of technological tools that contribute to the fluidity of development activities. The communication channels provided by these tools contribute to the participatory culture of software development, where requirements include getting involved, learning and co-producing the code. However, the choice of channels to support development is still an open question in the scientific community. Through an investigation of closed remote software teams from a large public educational institution in Brazil during the COVID-19 Pandemic, we identified their preferred communication channels and analyzed them to learn how to improve collaboration. We also drew some comparison with social developers and took a look at the challenges that these channels impose and solutions adopted to mitigate these problems.

Competency Management Systems are used to facilitate individual competencies management, both in companies and academia. Most systems use the term “competency management” based only on the technical skills of individuals. This way, the manager ends up not being sure of the accuracy of the technical skills of individuals, forming low-performance and poorly collaborative groups. This paper presents an open source web tool aimed at forming highly collaborative groups. The main contribution of this research is the innovative use of the Forgotten Curve concept and modeling for collaborative based grouping to support the design of work teams.

Social isolation and loneliness are problems faced by the elderly that might be aggravated due to quarantine during the coronavirus pandemic. This motivates a search demand for automatic technological solutions, such as chatbots, to address the problem, and establish a collaborative interaction between elderly and conversational agents. We analyze the use of conversational agents and the main types of personalization used during interacting with the elderly. We mapped 53 papers in 5 sources to answer 5 research questions that guided our study. Our findings show these solutions may have a positive impact on elderly's life.

It is our pleasure to welcome you to this PACMHCI GROUP issue. For over 25 years, the GROUP research community has supported the development of robust scholarship at the intersection of Computer Supported Cooperative Work, Human Computer Interaction, Computer Supported Collaborative Learning and Socio-Technical Studies. This volume is the latest product of our efforts to validate and integrate the strong work happening within this broadly-conceived community. We hope that the range of papers presented here reflects our intent to be international, interdisciplinary, and inclusive, both in our organization of the review process as well as within the set of accepted papers. This PACMHCI journal volume in the GROUP series features studies of collaboration in multiple settings, including social media, online communities and game development. It also showcases architectures and frameworks for collaboration, and sociotechnical studies in domains ranging from sports to fake news. As Editors, we are particularly pleased to continue the legacy of presenting and disseminating the new and exciting work being developed in our community. We come together every two years to creatively display and build off of ideas from a wide range of disciplinary and topical areas including computer science, design, engineering, information science, management science, sociology, work and labor studies, and values-in-design, among others.

Collaboration is an important process for combining the potential and knowledge of experts in organizations. To support the transfer of knowledge, we highlight the role of teams in disseminating knowledge and in preparing new specialists. After all, the absence of a strategy to avoid the loss of a specialist's knowledge, by retirement or unexpected departure, may correspond to a reduction of competitive advantage for the organization. Therefore, this work proposes providing managers with mechanisms that facilitate the formation of new specialists through a behavioral approach. For this, collaborative observations, based on the use of the eMOC tool - Mapping Expert Behavior Through Observation, can intensify collaboration productivity, coordinate activities and improve interaction among group members. The result of the work consists in presenting a set of cognitive barriers and thus promoting the acceleration of the transfer of knowledge, generating profit and value for the organizations.

The analysis of data collected through GPS and/or smartphone logs allows the identification of human mobility patterns. This identification has been the subject of several surveys in recent years. However, certain factors can change these patterns, such as special events of great magnitude (e.g., the Olympic Games) or environmental events (e.g., an increase in air pollutants.) In this paper, we present an investigation into the impact of these events on these patterns. We analyze urban mobility data collected in Beijing between April 2009 and October 2012, we use Artificial Intelligence techniques to identify mobility patterns in these data and then check for changes in patterns during the aforementioned events. The analysis of these data has become an important tool to extract insights and to support the planning of the urban displacement in the day to day and in special events.

Online collaborative environments, such as Reddit, have led to the rise of open and self-organizing communities. However, discussions in these places are often scattered, redundant and voluminous, and do not fully support the deliberation flow. Argument polarization, groupthink, subject bubbles, subgroup dominance and lurking are commonly identified behaviors that negatively impact the quality of online discussions. In addition, open environments allow the congregation of homogeneous groups that feed on pre-existing ideas, creating discussion bubbles and contributing to users taking extreme positions on issues. Evidence from Reddit discussions suggest that, while online collaborative environments provide powerful support for the creation and dissemination of new content, their ability to support efficient debate and idea contribution is questionable. Often, debates include hundreds or thousands of people with heterogeneous behaviors concerning post frequency, size, wording and polarity. Joining an ongoing online debate requires understanding what has happened so far. Many studies have focused on summarizing the discussion content as a way to overcome the message volume, dispersion and redundancy. However, to fully understand an online debate, it is also important to perceive the group's dynamics. The identification of dysfunctional behaviors contributes to new users' sensemaking of the debate as a whole and supports their decision whether it is worth joining the group. Additionally, understanding the group dynamics improves participants behavior and the overall discussion. We analyze the use of content summarization and behavior indicators to understand a discussion.

It is our pleasure to welcome you to the first PACMHCI GROUP issue. For over 25 years, the GROUP research community has supported the development of robust scholarship at the intersection of Computer Supported Cooperative Work, Human Computer Interaction, Computer Supported Collaborative Learning and Socio-Technical Studies. This volume is the latest product of our efforts to validate and integrate the strong work happening within this broadly-conceived community. We hope that the range of papers presented here reflects our intent to be international, interdisciplinary, and inclusive, both in our organization of the review process as well as within the set of accepted papers. This first PACM journal volume in the GROUP series features studies of collaboration in multiple settings, including social networks, editing systems, and mixed reality. It also showcases sociotechnical studies in domains ranging from outdoor activities to scientific projects.

Knowledge transfer in organizations is a decisive advantage in the competitive market. When it is done effectively, it prevents companies from losing relevant information when experts leave or are disconnected for any reason. Considering that beliefs and collaboration are important to support the knowledge transfer process, this paper investigates the main cognitive barriers and beliefs that may prevent knowledge transfer in organizational environments. This work maps the knowledge transfer processes found in organizations in order to elicit the barriers faced by experts as well as their beliefs, and how they affect the group during complex decision activities. These barriers are part of the Conceptual Framework for Identifying Barriers of Knowledge Transfer between Experts (KTAe).

Experts apply their intuition in different activities within organizations (e.g. decision making, impact projection). Thus, when an expert leaves the organization, it loses the expert's intuition as well as his/her expertise. In this context, the organization spends resources (e.g. time, money) to replace and train a new expert. We propose to use a team instead of an individual as a way to accelerate the process of training and replacing the expert. For the team to be able to perform actions using a rationale similar to the individual expert's, a process to develop intuition in a group is proposed. The application of the process in a real scenario provides initial evidences about its viability.

Data Science has assembled researchers from multiple fields, such as computer science and marketing, to develop methods, algorithms and techniques to discover knowledge from large amounts of data and solve complex problems. The main focus is to find patterns hidden within the data. In general, these patterns must appear frequently to be learned using machine algorithms and reflect the standard, even if tacit, knowledge that guides decisions in the world. The uncovered knowledge may also reflect societal prejudice that will be enforced by the machines. Nevertheless, in many domains, specially when the negative impacts of bad decisions have a high cost, a few instances of patterns within the dataset suffice to warrant further investigation. Considering that decisions will be based on knowledge learned from the data, the challenge lies in determining when fewer appearances of a sequence of events are more important than very frequent patterns. In this context, we have devised an approach that uses a domain ontology to boost these infrequent, but relevant, events. In addition to guiding the search for relevant knowledge, the ontology helps users accept results and further investigate the data. It enables users to create data subsets and views, deriving new attributes from existing ones, guiding the data mining process, and providing a background layer from which even not so frequent patterns stand out and become meaningful. In this paper, we present our ontology-based data discovery approach, a system developed according to it, and preliminary results of a real life application in the oil production domain.

Emergency command and control centers (CC) are integrated facilities to assist and handle crisis situations. In these CCs, operators suffer both with information shortage and overload. This paper focuses on the information overload problem. Operators often do not have adequate access to information that may be relevant in the decision-making process. In many CC centers, information is stored without filtering or refinement for later use, and its visualization is pre-defined at design time, offering no possibility of customization. To address the problems created by this lack of support, we propose an architecture to enable fast and easy selection from multiple information sources, and definition of appropriate visualizations. Following this architecture, we developed the “Emergency Dashboard”, a system that provides selection of a desired set of data and presents it, using configurable dashboards. We also provide collaborative mechanisms, given that, at times, it is not possible for a single user to handle such large datasets alone.

Crowdsourcing initiatives have been around for several years, and have been experiencing increasing adoption. Crowdsourced systems use the aggregation of efforts from multiple individuals to perform tasks or solve problems. The exponential growth of mobile computing has strengthened this approach, and several applications have appeared which enable individuals to provide information that can later be aggregated and reused. However, these applications are highly dependent on user contributions, and this can become a problem if not enough contributions are provided. Gamification strategies may help increase user contributions, through the creation of game-like interactions. This paper explores the design of gamification strategies to engage users to in real time data mapping.

One of the keywords in scientific research is collaboration, because closer relationships facilitates the exchange of knowledge, reduces the search time and resources required for its implementation. A collaborative work has less chance of error and has greater visibility. Scientific collaboration is also very encouraged by educational institutions and research funding agencies. However, scientific collaboration has still some challenges. Collaboration can vary according to several factors, such as the culture of the country, its economic situation, the area of study. The academic community has been studying collaboration issues in order to find better solutions to support and encourage collaboration among researchers. This work presents an environment that supports scientific collaboration, so that researchers can find potential research partners that work on closer topics and who seems to have common interests. It is intended to provide interaction between researchers in order to improve, streamline and enhance the results of scientific collaboration.

Design is a knowledge intensive activity, which depends on information to be undertaken. Previous designs, recent innovations, new techniques and materials must be taken into account when working on a new project, whatever its nature. Designing a plan is no different from designing a physical artifact, in that several constraints and requirements must be taken into account. Therefore, a collaborative approach that enables the refinement of plans through Public Engagement is of paramount importance to the design process. Many studies have been undertaken regarding the promotion of public participation in the planning process. In this situation, the population can be seen as an important source of information, because they experience the environments addressed by the proposed plan on a daily basis. However, the interaction between government and population presents cultural, political and technological challenges. In this paper, we present the Scenario-Based Collaboration Process (SBCP), which aims to Refine Plans through Public Engagement.

Emergency plans are fundamental for the speedy and effective response in disaster situations. Plans are often constructed by teams of experts, who apply their expertise to define response procedures, but lack part of location-specific knowledge that can be very relevant to make decisions during resp

People and organizations frequently need to recall past events that, for some reason, were not documented when they occurred. The successful reconstitution of past events depends on several variables, such as whether key people are still available to tell what they know. This is especially true in e

In major disasters it is common to install command posts close to the affected areas, in which a unified command, composed by people with different skills and goals, works together to resolve the situation. A major difficulty when supporting this command post is filtering the information that should be available and providing it at the right time to the right people. Through the application of Cognitive Task Analysis methods, we elicited how the work is really done and how the information is used by the command staff. Based on the results of this analysis, we present a set of requirements to support collaboration and information availability in this type of situation. We also present the initial development steps of a system prototype that will support the command post activities.

Scenarios are cornerstones of emergency plan development because they help emergency managers envision a hypothetical real situation that may occur in the future. Scenarios are composed of events, but this is not enough to reach their entire potential. The use of scenario variables along with CIA-ISM approach can be used to support the generation and interrelation of events, expanding thereby the analysis of a scenario with situations closest to reality. The more people involved in generation, the better is the result (scenario) reached. This first step of a work-in-progress in emergency plan generation aims at facilitating the collaborative scenario creation by emergency experts with the adoption of both, variables and the CIA-ISM.

Events that threatens the status quo must be appropriately managed in order to reduce risk and bring the situation back to normal in a speedy fashion. In general, there is a lack of up to date information in emergency situations, this could be alleviated through the use of contextual information generated by social media sites. In this paper, we propose a model for social media information collection, which could be used by official organizations as an alternative communication channel aimed at reducing the distance between society and the command room.

The following section of the proceedings includes papers presented in the context of ISMICK, the International Symposium for the Management of Industrial and Corporate Knowledge, held in conjunction with CSCWD in the framework of the activities of the IEEE SMC CSCWD Technical Committee. The paper summarizes the history of the symposium since its started in 1993, its original goal and the evolution of the different themes during the following years.

Design is a knowledge intensive activity, which depends on information to be undertaken. Previous designs, recent innovations, new techniques and materials must be taken into account when working on a new project, whatever its nature. Therefore, retrieving appropriate information is of paramount importance to the design process. Information retrieval is an established field, which has seen great progress in recent years. However, as the amount of unstructured information available increases, it becomes harder to retrieve the most relevant information using standard techniques and, frequently, users end up with results that don't fulfill their information needs. In this paper, we present an approach to improve retrieval using query expansion through ontologies allied to faceted navigation. We constructed the Ontoogle tool, which uses an ontology describing a domain to expand the query and define result ranking and presentation using Faceted Navigation. This proposal was evaluated in a case study and the results indicate that this approach contributes to increased satisfaction and confidence of users during their search process.

Good decisions require good understanding of the problem, the context and available options. This knowledge, in general, is dispersed through several, frequently unidentified, individuals. Collecting, filtering and putting together the relevant knowledge pieces to create solutions are difficult tasks. Many domains fit this scenario of distributed knowledge and creative solution's needs including governmental decision-making. Democratic governments constantly need to make sense of population demands to properly make decisions that attend overall will. However, low citizens participation in government decisions through democratic processes is an aspect that defies democracy itself. Large-scale citizens participation is difficult for the lack of technology to collect, filter and synthesize information. To deal with this problem, we developed an intelligent mobile network that allows broad citizens' participation and feasible government action. Previous research indicates cellular phones as the device with the broadest acceptance in all Brazilian social classes. Sending messages and voting by cell phones are broken barriers to consider the device as the natural input interface for citizen-government interaction. Additionally, message summarization and clusterization help government understand the demands. In this paper we present mGov, a system for eliciting and answering citizen demands in a participatory government structure. A prototype has been deployed with good results electronic

Collective Intelligence has received much attention in recent years, as organizations and businesses discover the power of crowds. With new technologies, such as blogs, twitter, wikis, photo sharing, collaborative tagging and social networking sites, people may create and disseminate content as never before. Many organizations are looking for ways to harness this power, which is being called collective intelligence. Additionally, many open-source initiatives have shown that it is possible to obtain high quality results from collectively produced work. In this paper, we consider the domain of emergency response. We have been working in this domain for a few years now, mostly focusing on small disaster relief groups and field command. One of the command problems is lack of information about the location where the disaster occurred. We believe some of this information could be crowdsourced, changing current practices and enabling more effective response to the situation. In this paper, we discuss the possibilities for crowdsourcing and present an initial proposal for harnessing collective intelligence for the fire department.

Meetings take a large amount of time, and frequently, meeting facilitators are called in to help steer the meeting in the right direction. Unfortunately, facilitation is still something of an art. To assist facilitators, the thinkLet technique has been proposed. Through the use of thinkLets, facilitators may plan the meeting beforehand, deciding which activities to perform at each point. However, designing a meeting is not an easy task, especially for novice facilitators. In this paper we present AgendBuilder, a system to assist the design of meetings, electronic or otherwise, using thinkLets and compatibility rules. The system helps facilitators select and link thinkLets together in and orderly fashion, designing the meeting in the process.

This paper presents observations from engineering ontology design sessions. We followed the entire design process of a group of engineers building an ontology for root cause analysis to be applied to petroleum pump failure diagnosis. The group was composed of experts with same background, who work in different locations of the same company. Even though they apparently share similar analysis methods, during meetings discrepancies emerged. Different points of view were argued and decisions were made at the end of each meeting. Given the final product acceptance problems, we look back at each meeting and attempt to identify the reasons that led to this unfortunate and common outcome. Understanding the problem enables us to draw a set of guidelines for empowering design meetings with simple technologic methods for improving group final artifact acceptance; therefore diminishing re-work. Our observations point towards (1) a correlation between designers' behavior and future artifact acceptance; (2) a set of actions that interrupts or brings back group attention; and (3) a metric for evaluating group design meetings. We believe our findings may guide software developers to tools for supporting group design.

Conferences are an important environment for academic knowledge exchange, and are valued by researchers as a good way to increase ones social network and meet potential collaborators. The question we investigate in this work is whether conferences actually influence collaboration between researchers, providing a starting point for joint work. To conduct this investigation, we used the DBLP database containing 640.977 distinct authors, 1.057.207 papers published in 21.354 journals and conferences spanning over more than 70 years. Using the concept of coauthorship as evidence of collaboration, we found that at best only 123.886 collaborations (from a total of 2.685.224, or 4,61%) could have emerged from the authors having met at a conference. This leads us to the conclusion that conference environments do not significantly promote collaboration among participants.

Scientific environments are known for being highly dynamic, subject to rapid evolution and demanding constant renewal and update. Additionally, science is a highly social arena. However, interpersonal collaboration and knowledge flow in scientific environments are usually more restricted. Collaboration is intense among small groups of people working on specific problems within a domain, but low between groups. As user profiling has been extensively used as a basis for recommendation, personalization and matchmaking systems, a better profile identification can improve interaction levels among researchers belonging to the same domain but working in different laboratories. Profiles may be constructed in two ways: either through explicit declaration by the user or through the observation of users’ actions. Many systems employ one approach to the exclusion of the other. We contend that a combined approach will yield better results, especially on scientific scenario, providing a mix of declared and inferred information. In this article, we present inference-based profiles in scientific environments (i-ProSE), an integrated system that dynamically creates and maintains scientific user profiles based both on declared information and on observed behaviour. The i-ProSE can be used to locate experts, deliver content, build communities, find collaborators for long-term projects or detect instantaneous opportunities for informal collaboration, which is presented with a short study case.

Decision-making involves choosing between one ore more alternatives, to achieve one or more goals. To support this process, there are decision support systems that employ different approaches, supporting groups or not. Generally, however, these systems do not have great flexibility; their users have to follow pre- established decision methods. This paper, after exposing some decision-making processes, describes a system, LaSca (from Large Scale), to support decisions in large-scale groups. This system, besides allowing effective achievement of the benefits of deciding in large groups through the proper structuring of the group, also allows its users to define themselves how this structuring will happen, based or not in the existing theories on the subject. So, in addition to facilitate the decision-making process, LaSca also allows its users to decide how to decide.

In many group work situations, meetings take up a reasonable amount of time. Several techniques have been introduced to streamline meetings and reach conclusions faster, many of which place an individual in the role of meeting facilitator. Facilitation involves intervening to produce desired results. Thus, a facilitator should be able to identify certain group dynamics and act accordingly. In this paper, we present and electronic idea generation system based on thinkLets dynamics and designed to measure group dynamics. This information is provided to the facilitator so he or she can decide whether or not to intervene.

Decision-making involves choosing between one ore more alternatives, to achieve one ore more goals. This process is composed of a series of steps, ranging from the identification of the problem itself, going trough the definition of criteria that characterize possible solutions to the problem, to the decision-making itself, in which a possible solution to the exposed problem is finally adopted. These steps can be performed by a single person or by a group of people. In group decisions, the solution choice at the end of the process must be effectively made by the group. To support this process, there are decision support systems that employ different approaches, supporting groups or not. This paper defines the decision-making process, presenting benefits of group decision and describes a system to support decisions in large scale groups. Participatory design in the BamPetro environment preservation project seeks to bring a large number of end users into the design process to take into account different perspectives, and help the group cope with multi-objective decisions. These decisions demand synergy among users of the environment preservation project, even though they represent different areas, competencies, political agendas and social interests.

In large communities, finding experts or partners can sometimes be hard. Normally, to find new partners one asks for recommendations from known acquaintances. This is a form of social navigation, whereby one searches for others in the network, who might fit required profiles. Time is an important factor in this type of situation, as needs, interests and expertise levels change rapidly. This paper presents a peer-to-peer system to assist social network navigation and search. It builds user profiles through an analysis of ongoing activities that takes time into account, and matches these to find opportunities for collaboration.

The SYMBAD (Similarity Based Agents for Design) system is a project developed to aid architectural design team members to share knowledge used in stand design. Stands play important role in promoting a company image, projects its presence and emphasizes the corporate identity in events and fairs. The processes and problems in stand projects are quite common and can be easily found in other design situations. From conceptual design to the construction of a final product, a stand project passes through many hands, each one adding bits and pieces until it is completed. The physical distance, difficulties in communication and cultural aspects make designers interact less than necessary, generating problems in all phases of the design activities. The SYMBAD project uses an ontology based framework to improve process awareness in a design activity. Based on the knowledge kept from previous projects and similarity aspects, agents can identify possibilities for reuse and provide information to help the design team to built new stands using past materials and ideas. Through a constructed ontology and collaborative features agents can produce global awareness, to facilitate later steps and optimize the process as a whole. In this paper we present some of the junctions and architecture aspects, as well as some of the implementations choices made for this project

Decentralized projects, where several organizations come together to work on a joint project, are quickly becoming the norm, due to increased networking and distribution. Subcontracting and consulting lead to the creation of more flexible, less hierarchical structures. In these networked structures, partners are often considered as peers and the group has the flexibility to restructure itself according to its needs. In this paper, we present a study of a decentralized software research, design and development group, and discuss its needs. A large amount of time is spent with articulation and maintaining partners up-to-date, which suggests that appropriate awareness support might facilitate the group's activities

With the increase in adoption of remote work groups and teams working in non-collocated environments, several opportunities for interaction have been lost. Contact with colleagues and discovery of opportunities, previously easy due to physical proximity, is made harder with distance. In this paper, we describe an agent-based framework to assist users in finding opportunities for interaction that may have been overlooked.

Stands have an important role in an event. They present the image of a company, project its presence and emphasize the corporate identity to all guests. Architecture companies, which design stands, involve several individuals. From conceptual design to the construction of a final product, a stand project passes through many hands, each one adding bits and pieces until it is completed. The physical distance, difficulties in communication and cultural aspects make workers interact less than necessary, generating problems in all phases of the design activities. Reuse of materials and ideas is less feasible as design complexity increases. The processes and problems in stand projects are quite common and can be easily found in other design situations. We present an agent framework to improve process awareness in an architecture company. Based on users activities and previous designs, agents identify possibilities for reuse and provide information to the conceptual designer so that their designs take construction difficulties and possibilities for reuse into account. The agents instrument the process to produce global awareness, to facilitate later steps and optimize the process as a whole. In this paper we present the agent architecture, as well as each agent's general functioning and reasoning rules.

When a novice needs help, often the best solution is to find a human expert who is capable of answering the novice's questions. But often, novices have difficulty characterizing their own questions and expertise and finding appropriate experts. Previous attempts to assist expertise location have provided matchmaking services, but leave the task of classifying knowledge and queries to be performed manually by the participants. We introduce Expert Finder, an agent that automatically classifies both novice and expert knowledge by autonomously analyzing documents created in the course of routine work. Expert Finder works in the domain of Java programming, where it relates a user's Java class usage to an independent domain model. User models are automatically generated that allow accurate matching of query to expert without either the novice or expert filling out skill questionnaires. Testing showed that automatically generated profiles matched well with experts' own evaluation of their skills, and we achieved a high rate of matching novice questions with appropriate experts.

Web browsing, like most of today's desktop applications, is usually a solitary activity.Other forms of media, such as watching television, are often done by groups of people, such as families or friends.What would it be like to do collaborative Web browsing?Could the computer provide assistance to group browsing by trying to help find mutual interests among the participants?Let's Browse is an experiment in building an agent to assist a group of people in browsing, by suggesting new material likely to be of common interest.It is built as an extension to the singleuser Web browsing agent Letizia.Let's Browse features automatic detection of the presence of users, automated "channel surfing" browsing, and dynamic display of the user profiles and explanation of recommendations.