[cfp][meetings] The 5th Workshop on Rebellious and Disobedient Agents in AI (RaD-AI) at AAMAS 2026

[Mirsky, Reuth]

Call for Papers

AAMAS-26 Workshop on Rebellious and Disobedient Agents in AI (RaD-AI)

May 25 or 26 (TBD), 2026; Paphos, Cyprus

https://sites.google.com/view/rad-ai/rad-ai

Should intelligent autonomous agents always obey human commands or instructions? We argue that, in some contexts, they should not. Most existing research on collaborative robots and agents assumes that a “good” agent is one that complies with the instructions it is given and works in a predictable manner under the consent of the human operator it serves (e.g., it should never deceive its operator). The goal of this workshop is to challenge this assumption and to rethink the desired abilities and responsibilities of collaborative agents. These include, for example, exhibiting behavior that attempts appropriate and harm-preventing non-compliance (e.g., safety constraints in autonomous vehicles or training LLMs to avoid potentially harmful or norm-violating output), among others. 

Truly collaborative agents should do more than simply comply with given commands. However, this deeper notion of collaboration raises questions. When and how should an artificial agent engage in non-compliance? Additionally, what sort of novel computational capabilities are required to enable agents to (1) interpret their operator’s objectives and societal expectations, (2) reason about these in the context of their environment, their own objectives and experience, and their teammates (and other agents) and (3) select appropriate response objectives, plans, and actions that may be considered, to outside observers, to be “rebellious” or “disobedient”? What distinguishes perceived “disobedience” from true disobedience? Would true disobedience from autonomous agents ever be desirable? 

We invite the participation of researchers interested in directly addressing Rebellious and Disobedient Agents in AI (RaD-AI), as well as submissions on RaD-AI and related topics relevant to designing, developing, demonstrating, or evaluating RaD-AI agents, including but not limited to:

• Intelligent Social Agents (including but not limited to: Goal Reasoning, Plan Recognition, Value Alignment, and Social Dilemmas)
• Human-Agent/Robot Interaction (including but not limited to: Human-agent Trust, Interruptions, Deception, Command Rejection, and Explainability) 
• Societal Impacts (including but not limited to: Legal and Ethical Reasoning, Liability, AI safety, and AI governance)

We are also interested in and encourage the submission of interdisciplinary work relevant to these and related topics. 

Paper Submissions:

Papers should be submitted through EasyChair: https://easychair.org/conferences/?conf=radai26

 

We accept submissions of the following types: regular research papers (up to 8 pages), short papers (up to 4 pages), position papers (up to 2 pages), and tool talks (up to 2 pages).

 

Papers must be in high-resolution PDF format, formatted for US Letter (8.5" x 11") paper, using Type 1 or TrueType fonts. Reviews are double-blind, and submissions must conform to the AAMAS-26 submission instructions found here:  

https://cyprusconferences.org/aamas2026/submission-instructions/.

Dates:

February 4th, 2026: Workshop Paper Submission Deadline

March 20, 2026: Acceptance Notifications Emailed to Authors

TBD: Workshop Registration Deadline

May 25 or 26, 2026 (TBD): Workshop

Organizing Committee (alphabetized):

David W. Aha, Navy Center for Applied Research in AI, Naval Research Laboratory

Gordon Briggs, Navy Center for Applied Research in AI, Naval Research Laboratory

Peta Masters, Department of Informatics, King’s College London

Reuth Mirsky, Computer Science Department, Tufts University (primary PoC: reuth....@tufts.edu)

Sarath Sreedharan, Computer Science Department, Colorado State University

Mor Vered, Department of Data Science & AI, Monash University

Ruchen (Puck) Wen, Colgate University