Exercise Advisor ("Laura") Images
Click on a thumbnail for an enlargment.


Exercise Advisor ("Laura") Video (26.5MB)
This is a self-extracting video player and video clip. 
  Just download the file to your hard disk and run it. 
  It will take a minute or two to uncompress. PC only. 


Relational Agents are computational artifacts designed to build and maintain long-term, social-emotional relationships with their users. Central to the notion of relationship is that it is a persistent construct, spanning multiple interactions, thus Relational Agents are explicitly designed to remember past history and manage future expectations in their interactions with users. 

Since face-to-face conversation is the primary context of relationship building for humans, this work focuses on relational agents as a specialized kind of embodied conversational agent, which are animated humanoid software agents that use speech, gaze, gesture, intonation and other nonverbal modalities to emulate the experience of human face-to-face conversation. 

The theory of relational agents includes a multidimensional model of human-computer relationships, a taxonomy of verbal and nonverbal strategies which can be used by a relational agent to move the relationship in a desired direction, and a dialogue planner which can plan conversational strategies that work towards the achievement of both task goals and relational goals.

Preliminary work in this area  focused on the use of small talk by a relational agent within the context of a service encounter, specifically  real estate sales. In an experiment conducted with 31 subjects in the summer of 2000 we found that the use of social dialog by an agent made some users (extroverts in particular) trust the agent more than when it only talked about the task being performed. Establishing such trust is essential as computers start to help people with important tasks such as improving their health or buying a home.

Recent work  demonstrated the ability of relational agents to establish and maintain relationships with people over a series of interactions. In this effort, the agent played the role of an exercise advisor designed to motivate  users to exercise more. One hundered subjects participated in a six-week study longitudinal study (four week intervention and two week follow up) to determine the efficacy of this agent. Results indicate that the agent was successful at creating and maintaining a trusting, caring relationship with users and increasing their desire to continue interacting with it.


Bickmore, T. "Relational Agents: Effecting Change through Human-Computer Relationships", MIT Ph.D. Thesis, February 2003 [PDF]

Bickmore, T. and Picard, R. "Subtle Expressivity by Relational Agents", CHI2003 Workshop on Subtle Expressivity for Characters and Robots, (to appear). [PDF]

Bickmore, T.,  Cassell, J.  "Social Dialogue with Embodied Conversational Agents" In J. van Kuppevelt, L. Dybkjaer, and N. Bernsen (eds.),Natural, Intelligent and Effective Interaction with Multimodal Dialogue Systems. New York: Kluwer Academic. (to appear)  [PDF]

Bickmore, T. "When Etiquette Really Matters: Relational Agents and Behavior Change", AAAI Fall Symposium on Etiquette for Human-Computer Work, 2002 [PDF]

Bickmore, T.  "Social Dialogue is Serious Business." CHI 2002 Workshop on Socially Adept Technologies. [PDF

Cassell, J. and Bickmore, T.  "Negotiated    Collusion: Modeling Social Language and its  Relationship Effects in Intelligent Agents" User  Modeling and Adaptive Interfaces, 12: 1-44. 2002 [PDF

Bickmore, T. and Cassell, J.  "Relational Agents: A Model and Implementation of Building User Trust."  ACM CHI 2001 Conference Proceedings, Seattle, Washington, 2001.  [PDF

Cassell, J. and Bickmore, T.  “External Manifestations of Trustworthiness in the Interface” Communications of the ACM 43(12). [Abstract] [PDF

Bickmore, T. and Cassell, J. "'How about this weather?' - Social Dialogue with Embodied Conversational Agents." Proceedings of the AAAI Fall Symposium on Socially Intelligent Agents. North Falmouth, MA, 2000. [PDF


  This material is based upon work supported by the National Science Foundation under Grant No. 0087768. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.