AI is making what used to be impossible possible and opening new opportunities to create, deliver, and capture value! It's driving transformational change and innovation.
The AAAI 2027 Human-AI Collaboration Workshop explores how advancing AI can create new human-AI capabilities and shared agency for tasks requiring both human judgment and AI.
The workshop will focus on designing more effective, responsible, and empowering partnerships between people and AI agents to drive new kinds of innovation that transform business outcomes and the human experience.
AI is shifting from a tool that answers queries to a collaborator that plans, acts, builds, engineers, and delegates in collaboration with people. This workshop advances the science of human-AI collaboration: how to structure, coordinate, evaluate, and govern joint work between humans and AI agents.
It brings together researchers from AI, Human-Computer Interaction, Human-Robot Interaction, Cognitive Science, Organizational Science, Strategic Management, Economics, and Law to build a shared research agenda.
Financial services and the legal profession are focal domains for human-AI interaction. They involve high-stakes, heavily regulated settings where decisions, evals, guardrails, shared context, and multi-agent coordination must align and work reliably at scale.
Advancing Human-AI Collaboration
The AAAI Human-AI Collaboration Workshop brings together researchers, practitioners, and students to explore how humans and AI can achieve goals neither can achieve alone by working together, sharing control, and making joint decisions. Our goal is to foster transdisciplinary dialogue and shape the future with responsible AI that works effectively with people to sense more, solve problems, seize opportunities, and transform what is possible.
Topics of interest include: trust, models of human-AI teaming, delegation, and mixed-initiative interaction; complementarity, over and under-reliance, and automation bias; communication, groundedness, and context sharing between humans and agents; evals, guardrails, oversight, escalation, and accountability for agentic AI action; multi-agent systems with humans in and on the loop; AI as thinking partner, agentic experts, and digital twins of organizations and markets; human training, evaluation of team performance, trust calibration, and deskilling; collaboration in financial services and other high-stakes domains (law, medicine, science, government), including regulatory and professional-responsibility implications.
Human-AI Evals and Guardrails
Evals and benchmarks are critically needed to design and measure human-AI performance, alignment, trust calibration, skill retention and deskilling, and long-term outcomes that move beyond single-turn accuracy toward longitudinal, ecologically valid evaluation.
High-Stakes Human-AI Collaboration
Human-AI collaboration in financial services, legal practice, medicine, engineering, science, and government, where professional duties, liability, and ethical obligations constrain how work may be shared with AI; implications of collaboration structures for law, professional responsibility, and regulation (e.g., supervision duties, disclosure of AI involvement, allocation of liability within human-AI teams
Human-AI Partners
AI systems as cognitive collaborators in reasoning, strategy, and decision-making (brainstorming, critique, scenario exploration, and augmentation of expert judgment) and the design choices that make a "thinking partner" enhance rather than substitute for human thought; digital twins of organizations.
Workshop Description and Focus
The AAAI 2027 Workshop on Human-AI Collaboration: Partnership, Complementarity, and Shared Agency examines the transition from AI as a query-answering tool to AI as an active collaborator. The workshop centers on systems that can plan, act, delegate, and coordinate alongside people over extended horizons, focusing on how humans and AI can jointly shape goals, strategies, and responsibilities in complex environments.
We highlight diverse domains where such collaboration already emerges: code writing with software developers, contract drafting with lawyers, co-authoring analyses with scientists, and coordinating tasks across teams and partnerships. Core research themes include modeling human-AI partnership, designing for complementarity rather than substitution, and understanding shared agency in decision-making and action. This is a research workshop associated with AAAI 2027; visitors are encouraged to explore the rest of the page for evolving details on topics, discussions, and materials as content becomes available.
By framing AI as a collaborator, the workshop invites contributions that investigate how to align capabilities, interfaces, and organizational practices so humans and AI can jointly achieve outcomes neither could achieve alone. Questions span theory, empirical studies, system design, and evaluation methods that capture nuance in partnership quality, division of labor, and mutual adaptation over time.
Through this lens, the workshop aims to connect researchers interested in trustworthy interactive systems, multi-agent coordination, human-centered design, and socio-technical perspectives on AI deployment. As the program, talks, and resources are finalized, additional information will appear across this website. Please continue through the page to learn more about the workshop’s themes, example use cases, and how the broader research agenda on human-AI collaboration is evolving within the AAAI 2027 community.
Important Dates
Workshop Submissions Due: Friday, Nov 20, 2026 AoE (Anywhere on Earth)
Notifications Sent to Authors (by organizers): Wednesday, Dec 2, 2026
AAAI 2027 Registration is OPEN
November 16: AAAI 2027 Workshop Registration Opens
December 15, 2026: Early Registration Ends
January 10, 2027: Deadline for Registration Refund Requests and Standard Registration Ends
Workshop Date: February 16-23, 2027 (TBC)
Organizing Team
The workshop's premise is that collaboration must be studied jointly across core AI and agents, human-centered interaction design, cognitive science, social and affective intelligence, organizational and strategic management, and law/professional practice. The workshop organizing team spans three continents, academia and industry, and has extensive experience organizing major conferences, workshops, and symposia.
Cutting-edge collaborative research brings together diverse experts, advanced technologies, and shared data to accelerate discovery and solve complex problems that no single discipline could tackle alone, e.g., markets and processes as shared workspaces in which humans and AI jointly model, simulate, and decide.
Mary-Anne Williams, Chief AI Scientist, Commonwealth Bank of Australia; Professor University of New South Wales; Codex Stanford University (Co-Chair and Main Contact).
Eric Horvitz, Chief Scientific Officer, Microsoft USA; Distinguished Fellow, Stanford Institute for Human-Centered Artificial Intelligence (HAI) (Co-Chair).
Hatice Gunes, Professor of Affective Intelligence and Robotics, University of Cambridge, UK
Roland Vogl, Co-Founder and Executive Director, CodeX Stanford Center for Legal Informatics, Stanford Law School.
Michael Genesereth, Professor of Computer Science; Director of the Stanford Logic Group and Co-Founder of CodeX, Stanford University.
Harry Surden, Associate Director, CodeX Stanford University; Hatfield Professor of Law, Law School, University of Colorado, Boulder.
Workshop Venue at the 41st Annual AAAI Conference on Artificial Intelligence 2027
AAAI-27 will be held at the Palais des congrès Montréal, Canada and the workshop will be held at some stage between Feb 16-23.
More information about local accommodations can be found here: AAAI-27 Accommodation and Transportation Information