{"id":1169836,"date":"2026-04-28T07:37:53","date_gmt":"2026-04-28T14:37:53","guid":{"rendered":"https:\/\/find.codeghost.online\/en-us\/research\/?post_type=msr-academic-program&#038;p=1169836"},"modified":"2026-06-24T10:53:09","modified_gmt":"2026-06-24T17:53:09","slug":"ai-and-the-new-future-of-work-cfp-spring-2026","status":"publish","type":"msr-academic-program","link":"https:\/\/find.codeghost.online\/en-us\/research\/academic-program\/ai-and-the-new-future-of-work-cfp-spring-2026\/","title":{"rendered":"AI and the New Future of Work CFP | Spring 2026"},"content":{"rendered":"\n\n<p class=\"wp-block-paragraph\">Call for Proposals, Spring 2026<\/p>\n\n\n\n\n\n\n<h2 id=\"about-the-program\" class=\"wp-block-heading\">About the program<\/h2>\n\n\n\n<div class=\"wp-block-columns is-layout-flex wp-container-core-columns-is-layout-8f761849 wp-block-columns-is-layout-flex\">\n<div class=\"wp-block-column is-layout-flow wp-block-column-is-layout-flow\" style=\"flex-basis:66.66%\">\n<p class=\"wp-block-paragraph\">Collaboration between people is a central mechanism in how work gets done, but AI doesn\u2019t yet work as well for teams as it does for individuals. Understanding how humans and AI can work better together in group settings is a critical area of the AI frontier. In this year\u2019s New Future of Work Call for Proposals, we\u2019re looking to support research that advances this frontier. Some of the research challenges we expect scientists to propose include (but are not limited to):<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Systems that help a team with AI dramatically outperform an individual with AI: AI systems that enable teams of people to robustly and substantially outperform appropriate comparison conditions (e.g., individuals with AI and teams without AI) on important tasks.<\/li>\n\n\n\n<li>Collaboration \u201carenas\u201d and world models: Developing simulation models that can help people train and test AI systems in collaborative settings.<\/li>\n\n\n\n<li>Envisioning a positive future of collaboration: Empirical and theoretical models of AI-empowered collaboration that have the potential to lead to positive outcomes for all stakeholders (individual participants, teams, organizations, society more broadly).<\/li>\n\n\n\n<li>Eliminating collaboration \u201cdrudgery\u201d: AI systems that reduce effort spent on aspects of teamwork that are empirically shown to be low value or aversive to participants (e.g., through prior literature, formative studies, or explicit measurement of perceived burden).<\/li>\n\n\n\n<li>Addressing the attention crisis in collaboration: AI systems that address communication overload at work by radically reducing low-value uses of attention in collaborative settings, allowing people to focus where their cognition and effort is most valuable.<\/li>\n\n\n\n<li>AI proactivity in collaborative settings: When and how should an AI proactively participate in a group conversation? How do we measure the cost of precision and recall errors in this context?<\/li>\n\n\n\n<li>Increasing fluidity in collaboration: How can people in an organization on-board (and step away from) a project more efficiently and effectively so that expertise can be brought to bear in new ways?<\/li>\n\n\n\n<li>Ensuring the work of a team benefits the team: Non-exploitative AI systems and corresponding incentive structures in which the value of any data created benefits the individuals and teams that created it.<\/li>\n\n\n\n<li>Leveraging behavioral science: AI systems and models that learn from well-defined negotiation scenarios, problem solving scenarios, conflict management scenarios and similar.<\/li>\n\n\n\n<li>New norms and etiquette of AI-empowered teams: Identifying successful norms and social frameworks for teams that have AI deeply embedded in their work.<\/li>\n\n\n\n<li>Strengthening core teamwork mechanisms: AI capabilities that can support and improve core teamwork mechanisms like grounding and common ground maintenance.<\/li>\n\n\n\n<li>Collaborative AI systems and approaches that focus on how AI can do different things than humans, not just replicate what humans do: AI\u2019s scalability, instantaneous ideation, externalized cognition and affordance of low-cost experimentation capabilities are all examples of underutilized but promising applications of AI rooted in its differences to humans, not its similarities.<\/li>\n<\/ul>\n\n\n\n<div class=\"wp-block-columns is-layout-flex wp-container-core-columns-is-layout-8f761849 wp-block-columns-is-layout-flex\">\n<div class=\"wp-block-column is-layout-flow wp-block-column-is-layout-flow\"><\/div>\n<\/div>\n<\/div>\n\n\n\n<div class=\"wp-block-column is-layout-flow wp-block-column-is-layout-flow\" style=\"flex-basis:33.33%\">\n<h4 id=\"key-dates\" class=\"wp-block-heading\">Key Dates<\/h4>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>April 28, 2026:<br><\/strong>Proposal period opened<\/li>\n\n\n\n<li><strong>May 25, 2026:<\/strong><br>Proposal period closed<\/li>\n\n\n\n<li><strong>June 15, 2026:<\/strong><br>Recipients notified<\/li>\n\n\n\n<li><strong>June 24, 2026:<\/strong><br><a href=\"https:\/\/find.codeghost.online\/en-us\/research\/academic-program\/ai-and-the-new-future-of-work-cfp-spring-2026\/recipients\/\">Recipients announced<\/a><\/li>\n<\/ul>\n<\/div>\n<\/div>\n\n\n\n<div class=\"wp-block-group is-layout-constrained wp-block-group-is-layout-constrained\">\n<p class=\"wp-block-paragraph\"><strong>Proposal requirements<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We\u2019re planning to support research groups from universities around the world to work on challenges like those listed above. We are aiming for a lightweight application process and proposals should be no more than 500 words plus references. Please include budget information in your 500 words.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Funding levels will be approximately $50K-75K USD, so proposals that articulate how funding from Microsoft will be used to complement or attract additional funds are encouraged. Given the speed of change in the AI world, we\u2019re also hoping to see work that can make material impact quickly.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The review cycle will be quick, and writing a proposal will be as well! Proposals are limited to one page, and should clearly speak to:<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li>The importance of the proposed research question.<\/li>\n\n\n\n<li>The potential impact of the outcome.<\/li>\n\n\n\n<li>The planned methodological approach.<\/li>\n\n\n\n<li>The qualification of the PI(s).<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">Proposals must be submitted from individuals from academic institutions.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Important Dates:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Proposals due: Monday, May 25th @ 11:59pm Pacific Time<\/li>\n\n\n\n<li>Decisions announced: Week of June 8th<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Eligibility<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">To be eligible for this RFP, your institution and proposal must meet the following requirements:&nbsp;<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Institutions must have access to the knowledge, resources, and skills necessary to carry out the proposed research.&nbsp;<\/li>\n\n\n\n<li>Institutions must be either an accredited or otherwise degree-granting university with non-profit status, or a research organization with non-profit status.&nbsp;<\/li>\n\n\n\n<li>Proposals that are incomplete or request funds more than the maximum award will be excluded from the selection process.&nbsp;<\/li>\n\n\n\n<li>The proposal budget must reflect your university\u2019s policies toward receiving unrestricted gifts and should emphasize allocation of funds toward completing the research proposed.<\/li>\n<\/ul>\n\n\n\n<div class=\"wp-block-buttons is-layout-flex wp-block-buttons-is-layout-flex\">\n<div class=\"wp-block-button\"><a data-bi-type=\"button\" class=\"wp-block-button__link wp-element-button\" href=\"https:\/\/find.codeghost.online\/en-us\/research\/project\/the-new-future-of-work\/\">Visit New Future of Work<\/a><\/div>\n\n\n\n<div class=\"wp-block-button\"><a data-bi-type=\"button\" class=\"wp-block-button__link wp-element-button\" href=\"mailto:newfutureofwork@microsoft.com\">Contact us<\/a><\/div>\n<\/div>\n<\/div>\n\n\n\n<div class=\"wp-block-group is-layout-constrained wp-block-group-is-layout-constrained\">\n<h3 id=\"conditions\" class=\"wp-block-heading\">Conditions<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">The following conditions apply to submissions and, as applicable if selected, your participation in the program. By submitting a proposal, you agree to the following:<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li>Microsoft may use your name and likeness to publicize your proposal (including proposal content) in connection with the AI and the New Future of Work in all media now known or later developed. For communication and community purposes participants&#8217; names and or email address will be visible to other participants on email invitations and in virtual meetings.<\/li>\n\n\n\n<li>Microsoft has no obligation to maintain the confidentiality of any submitted proposals.<\/li>\n\n\n\n<li>Proposals will not contain information that is considered proprietary, confidential, classified, restricted, or sensitive.<\/li>\n\n\n\n<li>The submission review process is internal to Microsoft. No feedback will be given to submitters. Microsoft may, however, reach out to submitters for clarifications on submitted proposals.<\/li>\n<\/ol>\n<\/div>\n\n\n\n<p class=\"wp-block-paragraph\"><em><strong>Note<\/strong>: Microsoft will award these as unrestricted gifts. Our accompanying gift letter will indicate this and that addresses overhead issues in most cases. However, some schools still do take a small overhead out and we have no ability to influence that beyond the letter we provide, as well as no way of knowing which schools will do what. If your school will take overhead, you need to consider that when building out your plans and include it within the max $75k your proposal requests.<\/em><\/p>\n\n\n\n\n\n<h2 id=\"2024-recipients\" class=\"wp-block-heading\">2026 Recipients<\/h2>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.cs.emory.edu\/~eugene\/\">Eugene Agichtein<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Emory University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>When to Ask AI or Colleagues: Discovering Team Conventions for AI-Supported Onboarding and Collaboration<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A new team member must learn both technical project details and social expectations: who to ask and how, and when one can rely on AI instead of seeking help from more experienced colleagues. This project investigates a key research question in AI-assisted collaboration: how to characterize team social expectations for seeking help, so that questions can be automatically routed to AI assistants or the right team members. Specifically, we aim to discover implicit expectations for communication, and to identify team experts and facilitators, by mining Teams chats, meeting transcripts, PR discussions, and commit and revision histories, with associated question engagement and community reactions. As a proof of concept and a platform for user studies, we will develop a TeamDispatch prototype that predicts if a question can be answered by an AI assistant, or suggests team members or channels to contact, utilizing user and team feedback to evaluate and refine performance. The project will produce new insights and question routing models that adapt to team expectations, changes in user needs, and advances in AI capabilities, to support effective team onboarding and collaboration.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/mattbrehmer.ca\/\">Matthew Brehmer<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Waterloo<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Evaluating the Utility of Proactive Glanceable Visualization Drawn from Organizational Knowledge Bases Delivered in Meeting Side Channels<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Synchronous conversations grounded in data are omnipresent in organizations, however the data assets grounding these meetings are too often fragmented, stored in heterogeneous formats and distributed across applications, from business intelligence dashboards to spreadsheets and static charts or tables on presentation slides. As a result, people laboriously search for, curate, and synthesize these assets before meetings, or they manually retrieve data assets during meetings, disrupting to the flow of conversation and imposing a context switch by redirecting attendees to data assets in other applications. Acknowledging that the precise information needs of meetings are difficult to predict, we are developing a proactive, mixed-initiative approach built on a multi-agent large language model (LLM)-based pipeline integrating retrieval-augmented generation (RAG). Using the real-time transcribed utterances of meeting participants, it delivers (via meeting side-channels) succinct and ephemeral glanceable summaries and information graphics drawn from an organizational knowledge base. This award will allow us to rigorously evaluate our approach\u2019s impact on workplace communication dynamics, such as in a longitudinal deployment study.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/hci.csail.mit.edu\/#faculty\">Zana Bu\u00e7inca<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Massachusetts Institute of Technology<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Beyond Autonomous Agents: Cultivating Metacognition and Sociality in Teams<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Current AI agents are optimized for individual productivity, a paradigm that creates a productivity trap: tasks get faster and cheaper, but the collaborative processes that produce higher-quality outputs \u2014 shared deliberation, mutual critique, diverse perspectives \u2014 are bypassed and eventually atrophy. Workers are left less skilled, more isolated, and less rewarded. We argue instead for designing long-horizon human-AI partnerships that promote collective flourishing: initiating a virtuous cycle where AI renews human expertise within teams, enables workers to identify new, more valuable forms of work, and produces outputs of greater depth and richness than any individual-AI pairing can achieve.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">To realize this vision, we investigate how AI can cultivate the most distinctive aspects of human labor in group settings: our collective capacities for critical thinking, reflection, and social connection. We pursue three directions: (1) collective metacognitive scaffolding, designing AI interactions that prompt teams to jointly monitor strategies, surface knowledge gaps, and articulate shared reasoning; (2) social priors for RL, using ethnographic observation of real workplace interactions to train models that understand the tacit norms of mentorship, perspective-taking, and constructive intervention within groups; and (3) organizational innovation, investigating how socially-aware agents can help teams invent new collaborative structures that humans define and implement themselves.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Mitchell Gordon, Arvind Satyanarayan, Graham Jones, and Susan S. Silbey, Massachusetts Institute of Technology<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/goizueta.emory.edu\/faculty\/profiles\/hancheng-cao\">Hancheng Cao<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Emory University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>TeamFlow: AI Facilitators for Attention Curation in Team Collaboration<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We propose to study and build AI agent that helps teams use their collective attention more effectively in collaborative work. Rather than helping teams process ever more information, we focus on reducing cognitive waste: situations where teams spend time and effort on low-value, redundant, or poorly allocated communication. The project will examine how AI can help teams identify which issues deserve attention, which can be delegated or deferred, and where human judgment is most needed. Our goal is to improve coordination, decision-making, and worker agency in AI-augmented teamwork.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Hong Shen and Qing Xiao, Carnegie Mellon University<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/lauradabbish.com\">Laura Dabbish<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Carnegie Mellon University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>TeamCompass: AI Facilitation for Collective Intelligence<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">TeamCompass investigates how AI can strengthen collective intelligence in human teams. Drawing on theories from organizational psychology, we are developing AI systems that facilitate team processes rather than simply assist individual users. Using AI red-teaming as a testbed, we will study how AI can help teams coordinate attention, distribute expertise, maintain shared awareness, and sustain diverse exploration when solving complex problems. The project will contribute new design principles, open-source tools, and empirical evidence about when and how AI can improve collaborative performance. Ultimately, this work aims to inform the next generation of AI systems that support teamwork in domains ranging from software engineering and cybersecurity to scientific discovery and organizational decision-making.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Laura Dabbish, Hong Shen, and Alice Zhang, Carnegie Mellon University <\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"http:\/\/www.manuelhoffmann.org\">Manuel Hoffmann<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of California, Irvine<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>When Do AI-Enabled Teams Outperform AI-Enabled Individuals?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Generative AI is transforming how individuals work, but organizations are now deploying it across entire teams, often without knowing whether that investment pays off beyond what individuals could achieve on their own. This project tackles that question directly: under what conditions do AI-enabled teams outperform equally AI-equipped individuals? We study this in a large enterprise setting where AI collaboration tools are being rolled out across dozens of teams, allowing us to rigorously compare performance before and after adoption. Our approach combines hard performance metrics (such as speed, quality, and rework rates) with in-depth fieldwork to understand how teams build, or fail to build, shared routines around AI. The goal is to produce both rigorous academic findings and a practical playbook that organizations can use to make smarter decisions about deploying AI at scale.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Duncan Houldsworth, University of California, Irvine<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/politicalscience.stanford.edu\/people\/simone-paci\">Simone Paci<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Stanford University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Building the AI-Native Research Lab: How Agentic Systems Aid Controlled Delegation<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This project develops a shared infrastructure to help academic research teams use AI more ambitiously without losing substantive human control over research decisions. The central challenge is that AI can make research assistants appear able to take on more complex tasks, while also increasing the risk that theoretical choices, measurement decisions, or evidence interpretation drift away from the project\u2019s intellectual goals. The proposed solution combines standardized AI-assisted workflows with clear escalation practices, so that routine work can move faster while judgment calls are surfaced to the appropriate human decision-maker. The project will study how research teams use this infrastructure through workflow records, documentation trails, and structured feedback from faculty and research assistants. The analysis will assess whether this approach improves delegation, research progress, documentation quality, and accountability while reducing preventable rework and coordination burdens.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.kellogg.northwestern.edu\/faculty\/uzzi\/htm\/index.htm\">Brian Uzzi<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Northwestern University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Seeing Teams Through Multimodal AI<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For AI to become an effective collaborator, it must first learn what effective collaboration looks like. This project will develop a model of collaboration using more than 2,500 hours of real-world team recordings from biomedical research teams, B2B sales calls, and entrepreneurial pitch interactions. By combining team science with multimodal AI, we will identify fine-grained interaction trajectories and examine which trajectories predict successful outcomes across settings. The project will produce a validated model of collaboration that can inform future AI systems designed to support early warning signals, process-aware coaching, adaptive team design, and reflective learning in the future of work.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Evey Huang, Northwestern University <\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/marynel.net\/\"><span class=\"sr-only\"> (opens in new tab)<\/span><\/a><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.bing.com\/ck\/a?!&&p=6d2b7068f053624109d9536340dfaf2a800a885e3c00960ca6592498b1e9f0b6JmltdHM9MTc4MTgyNzIwMA&ptn=3&ver=2&hsh=4&fclid=35a6fb3c-85ee-6862-223b-ed41849a696e&psq=marynel+vazques&u=a1aHR0cHM6Ly9tYXJ5bmVsLm5ldC8&ntb=1\" target=\"_blank\" rel=\"noopener noreferrer\">Marynel V\u00e1zquez<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Yale University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Adapting AI Agent Proactivity During Group Collaboration<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">How proactive an AI agent should be in a group is often group-dependent. An AI that intervenes too often can become disruptive, while one that stays silent could miss an opportunity to be helpful. How much is too much proactivity? This project frames this question as a preference learning task, where an AI agent learns a group&#8217;s proactivity preferences in real time, based on explicit and implicit signals from users. Experiments will focus on a creative storytelling task, where an AI agent supports pairs of users.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Etiosa Omeike, Yale University <\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"http:\/\/cs.cmu.edu\/~sherryw\">Sherry Wu<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Carnegie Mellon University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>A Collaboration Arena for AI Agents That Discover Hidden Team Needs<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">AI agents can help teams coordinate complex plans, but only if they account for the preferences stakeholders never state outright &#8212; constraints that are often latent, conflicting, and revealed only through interaction. This project develops MTP-Arena, an evaluation environment that tests whether agents can surface these hidden needs through dialogue and produce plans a full team will accept, rather than ones that merely satisfy the stated constraints. To make this scalable, we also build MTP-Forge, a pipeline that converts existing planning benchmarks into interactive multi-stakeholder settings. Together, they give the research community a concrete way to measure how well AI collaborators navigate the preference discovery at the heart of group decision-making.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Yilin Zhang and Xinran Zhao, Carnegie Mellon University<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/cs.stanford.edu\/~diyiy\/index.html\">Diyi Yang<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Stanford University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>From Individual to Teams: AI Agents for Empowering Human Teams<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">AI agents&#8217; adoption becomes common in organizations across domains, however, most agents are designed and optimized for single-user use cases. In the real-world human organizations, can agents help reduce the coordination overhead, misaligned goals, and breakdowns in common ground? Our recent CooperBench results show frontier coding agents lose roughly 50% of their performance the moment they must coordinate. The same coordination gap appears in human teams, but we lack the empirical foundation to design AI that closes it. This proposal pursues three connected questions: 1. Can we identify human cooperative\/uncooperative behaviors and the motivations behind them? 2. Can we predict human team work efficiency through simulation? 3. Can we build coordination agents that work alongside humans to improve the work efficiency and reduce the uncooperative behaviors?<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Hao Zhu and Yijia Shao, Stanford University<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/jieyuz.net\">Jieyu Zhao<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Southern California<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>From Personal Copilots to Team Coordination: Simulating Coordination Breakdown in Interdependent Knowledge Work<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Today\u2019s most deployed AI assistants optimize single-user interactions even though much knowledge work is interdependent. This creats a frontier gap: AI supports individuals better than teams. We argue that team productivity is not simply the sum of individual productivity, because locally optimal actions can produce delays, duplicated effort, and unresolved dependencies. This project builds TeamArena to simulate such coordination breakdowns and TeamOpt as a privacy-preserving coordination layer that helps personal agents coordinate around goals, ownership, blockers, and dependencies.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Yue Wang and Linxin Song, University of Southern California<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h2 id=\"2024-recipients\" class=\"wp-block-heading\">2024 Recipients<\/h2>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\">Fernando Diaz<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Carnegie Mellon University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Generative Tip of the Tongue Retrieval Support<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Tip of the tongue retrieval refers to searching for an item a user has seen before but whose name they do not recall.&nbsp; Although generative AI has led to advancements in how we search for information, tip of the tongue retrieval remains difficult due to the imprecise, incorrect, and contextual information provided by searchers.&nbsp; We propose a LLM-based simulation framework for developing interactive tip of the tongue retrieval systems.&nbsp; By adopting a retrieval-based evaluation and optimization architecture, this framework can be applied in enterprise settings without encoding confidential information in model parameters.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): To Eun Kim, Carnegie Mellon University<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"varol-akman-1\" class=\"wp-block-heading\">Sharon Ferguson<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Waterloo<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>A Novel AI-Powered System for Building Shared Understanding in Teams<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We envision human-AI collaboration that supports human-human collaboration by characterizing, analyzing and providing suggestions for collaboration grounded in team science literature. Shared understanding, or the extent to which team members are \u201con the same page\u201d about their goals, processes, and interactions, is important for efficient and enjoyable teamwork. Yet, lapses in shared understanding are easy for teams to miss. We aim to use traditional ML algorithms combined with LLMs to measure shared understanding based on a team\u2019s chat messages or video calls. The system will identify lapses in Shared Understanding in real-time and ask clarifying questions to help teams get back on the same page. By encouraging team members to reflect on their understanding, proactively identifying misunderstandings, and prompting perspective-taking, this system will bring the depth of team science literature to everyday teams.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Sirisha Rambhatla, University of Waterloo; Alison Olechowski, University of Toronto<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Gary Hsieh<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Washington<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Accelerating Research Translation into Design Practice Using Generative AI<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Valuable insights embedded in scientific publications are rarely translated into formats that design practitioners can easily consume and apply in their work. These research-to-practice gaps impede the diffusion of innovation and undermine evidence-based decision making. In this work, we explore the potential of generative AI to convert academic findings into a designer-friendly format. We will examine how to design these translational artifacts so that they are credible and are appropriately tailored to individual designers\u2019 needs.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Lucy L. Wang, University of Washington<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Harmanpreet Kaur<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Minnesota<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Using Adversarial Design to Support Appropriate Reliance on Generative AI Outputs<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The value of integrating generative AI into applications is limited by whether people are able to appropriately rely on its outputs\u2014while over-reliance can result in potentially harmful content being adopted, under-reliance prevents people from taking advantage of AI assistance. This project proposes adversarial design as an approach for supporting appropriate reliance on generative AI outputs, and tests its efficacy using LLM-based assistants for tasks like coding, search, and writing. For instance, we will introduce adversarial outputs in syntactic and semantic elements of the generated code, as well as via a chat interface that acts as a \u201c<a href=\"https:\/\/find.codeghost.online\/en-us\/research\/wp-content\/uploads\/2023\/12\/NFWReport2023_v5.pdf#page=9\" target=\"_blank\" rel=\"noreferrer noopener\">provocateur<\/a>.\u201d With this approach, we aim to reduce immediate acceptance or rejection of AI outputs, and nudge people to deliberate about them instead.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Rui Zhang<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">The Pennsylvania State University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Improving Productivity by Gradient-based Prompt Optimization for LLMs<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This proposal aims to develop a novel gradient-based prompt optimization technique to unlock the prowess of LLMs on complex and challenging tasks to elevate productivity. The novelty of our method lies in our design of a white-box prompt optimization algorithm over open-source language models by leveraging both the gradient of task accuracy and language model decoding probability. This leads us to automatically and efficiently create effective, interpretable, and transferable prompts on a variety of small and large language models on complex and diverse tasks. Our research advances prompt optimization by integrating gradients and LLM probabilities, creating long-term impacts on white-box optimization research to close the gap between small and large LMs.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h2 id=\"2023-recipients\" class=\"wp-block-heading\">2023 Recipients<\/h2>\n\n\n\n<h4 id=\"firstname-lastname\" class=\"wp-block-heading\">Eytan Adar<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Michigan, Ann Arbor<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Codes of Generative Conduct: Collaboratively Authoring and Testing Policies to Govern LLMs<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">When we write, we are guided not only by our intended audience but also by the constraints set by our communities\u2014the standards of our workplace, legal frameworks, and industry norms. Even sub-communities naturally develop their variants of these rules. For large language models to better support generative writing, they should adhere to the same set of community-set standards and rules. In our proposed work, we tackle how groups can collaboratively construct, test, and use writing standards to guide generative applications.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Eric Gilbert, University of Michigan, Ann Arbor<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"varol-akman-1\" class=\"wp-block-heading\">Varol Akman<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">\u0130HSAN DO\u011eRAMACI B\u0130LKENT \u00dcN\u0130VERS\u0130TES\u0130<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>ChatGPT as a Stage Manager in Large-Scale Institutional Practices<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Large-scale institutional structures are characterized by the collective action of numerous employees at different levels working towards common goals. Decision-making in these structures is complex due to the absence of egalitarianism and the presence of an authority hierarchy. Group agency emerges within these structures, where multiple agents work together based on joint commitments to advocate for shared goals. The project at hand aims to investigate decision-making and its consequences in such institutional structures using ChatGPT.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Danqi Chen<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Princeton University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Grounding Large Language Models with Internal Retrieval Corpora<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This proposal tackles the research problem of augmenting large language models (LLMs) with internal retrieval corpora, so we can enable LLMs to generate text that conforms to up-to-date and domain-specific information without additional re-training. The proposal aims to address two technical challenges: a) how to incorporate a large number of (and potentially noisy) retrieved tokens in a limited context window; b) How to train high-quality retrievers for specific domains for better integration with black-box LLMs.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Noshir Contractor<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Northwestern University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Deciphering the Structural Signatures of High-Performing Human-AI Teams<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This research project explores the dynamics of human-AI collaboration in problem-solving and creative thinking tasks. We will extend prior Human-Autonomy Teaming research that studies human-AI teams with the Wizard-of-Oz methodology\u2019s help by replacing a human confederate with an LLM-powered AI teammate. The experiment involves manipulating task types and AI teammate functions to examine how people orient themselves toward intelligent machine teammates and how technology can be designed to be a collaborator, not a chatbot. Participants will communicate in a group chat environment while conducting tasks, and their experiences, performance, and social interactions will be surveyed, analyzed, and compared to identify the mechanisms of high-performing Human-AI teams.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Philip Guo<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">UC San Diego<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Beyond just programming: using an LLM\u2019s knowledge about the human world to improve human-AI collaboration for data science<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We plan to build an LLM-based data science assistant that reasons about both the content of the code that it generates and the real-world domain that the requested analysis is about. Combining knowledge about both code and the human world can hopefully enable our LLM to perform more rigorous data analyses than current coding assistants like GitHub Copilot. The goal of this project is to advance the new future of work by empowering everyone across the workplace \u2014 from field technicians to customer support representatives to executives \u2014 to directly analyze data that matters most for their jobs. More broadly, we want to use AI to enable everyone to analyze the data that is most meaningful and relevant to them, even if they are not data science or programming experts.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">John Horton<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Massachusetts Institute of Technology<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Understanding the Effects of Github Copilot on Worker Productivity and Skill Evolution in the Online Labor Market<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Our project aims to understand the impact of Github Copilot on software engineers and other technical professionals\u2019 productivity, skill adoption, and labor market outcomes in a real labor market. LLMs have been shown to affect worker productivity on researcher-assigned tasks, but less is known about how workers use such tools in their real-life work, or what happens in equilibrium as more workers adopt them. We plan to use data from a large online labor market merged with survey data where workers report when they began using Copilot to understand how this tool impacted workers productivity, wages, and skill upgrading.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Co-author(s): Apostolos Filipas, Fordham University; Emma van Inwegen, MIT<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Anastasios (Tasos) Kyrillidis<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Rice University<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Efficient new-task adaptation in the era of Transformers and Federated Learning<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Starting from traditional Transformer models, one of the goals of this project is to introduce efficient ways to i) adapt to new tasks without re-training the models from scratch; ii) combine existing trained models in a meaningful way, extending MoE modules beyond basic feed-forward layer splitting; and iii) consider novel federated learning scenarios based on Transformer-based models, where computational-\/communication- bottlenecks require novel transformer decompositions, beyond sparse MoEs. Prior advances from this collaboration (results on asynchronous federated learning scenarios, new uses of a mixture of experts for zero-shot personalization in federated learning, and novel transformer-model decomposition for efficient distributed computing, respectively) will be extended based on the above directions.&nbsp;<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Matthew Lease<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Texas at Austin<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Providing Natural Language Decision-Support via Large Language Models<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Large language models (LLMs) are transforming the way people seek, explore, and assess information for decision making. By attending to how people naturally interact with each other, LLM-based decision-support tooling can strengthen both the quality of support offered and the accompanying user experience. Because LLMs are fallible and can hallucinate incorrect information, LLM interactions should also be designed to handle such fallibility in stride, akin to how we interact with one another around normal human fallibility. This project will explore such opportunities and challenges in designing ways to provide LLM-based decision support that is innovative, intuitive, and effective.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Mor Naaman<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">Cornell Tech<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Impact of Language Models on Writing and its Outcomes in the Workplace<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Use of autocomplete features and other language and text suggestions powered by large language models can shift people\u2019s writing, influence their thinking, and may lead to different outcomes for the person exposed to them. Using large-scale online experiments, this project aims to understand the potential of such AI-based products used in workplace settings to result in these different outcomes. In particular, we aim to understand the disparate effect these products may have on people from different backgrounds who may have different language styles. We will further investigate strategies to mitigate any negative outcomes.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Hamed Zamani<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Massachusetts Amherst<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Improving Productivity by Personalizing Large Language Models<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Large language models (LLMs) have recently revolutionized natural language processing, including a number of user-facing applications that largely impact users\u2019 productivity, such as QA, writing assistants, and task management. In user-facing applications, it is widely accepted that users have different needs and behave differently. This suggests the importance of personalization in these applications. The goal of this project is to study different approaches for personalizing LLMs and their potential impact on user\u2019s satisfaction and productivity.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n\n\n\n<h4 id=\"name\" class=\"wp-block-heading\">Amy Zhang<\/h4>\n\n\n\n<p class=\"wp-block-paragraph\">University of Washington<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Interactive Personalized Information Artifacts to Summarize Team Communication<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We examine how to design team-AI systems embedded into teams\u2019 existing communication spaces that can use LLMs to leverage rich prior communication data to support team goals and activities. Information artifacts that convey information from prior communication logs should not only summarize content but also support additional team goals such as sharing out different information to different audiences and serving as a springboard for follow-on discussion. We aim to use LLMs in a novel system for generating interactive summary artifacts that live within group chat and that summarize recorded video meetings. These artifacts will use LLMs to 1) personalize the content of the summary to different team members based on their conversation history, such as by highlighting specific points for follow-on discussion, 2) enable users to interactively expand the summary and dive into the context of the original conversation, and 3) allow users to customize an artifact in order to share it with different audiences.<\/p>\n\n\n","protected":false},"featured_media":1030995,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr_hide_image_in_river":0,"footnotes":""},"msr-opportunity-type":[155533],"msr-region":[256048],"msr-locale":[268875],"msr-program-audience":[],"msr-post-option":[],"msr-impact-theme":[],"class_list":["post-1169836","msr-academic-program","type-msr-academic-program","status-publish","has-post-thumbnail","hentry","msr-opportunity-type-grants-and-fellowships","msr-region-global","msr-locale-en_us"],"msr_description":"","msr_social_media":[],"related-researchers":[],"tab-content":[],"msr_impact_theme":[],"_links":{"self":[{"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-academic-program\/1169836","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-academic-program"}],"about":[{"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-academic-program"}],"version-history":[{"count":22,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-academic-program\/1169836\/revisions"}],"predecessor-version":[{"id":1176840,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-academic-program\/1169836\/revisions\/1176840"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/media\/1030995"}],"wp:attachment":[{"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1169836"}],"wp:term":[{"taxonomy":"msr-opportunity-type","embeddable":true,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-opportunity-type?post=1169836"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=1169836"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1169836"},{"taxonomy":"msr-program-audience","embeddable":true,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-program-audience?post=1169836"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1169836"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/find.codeghost.online\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1169836"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}