Groupsheets for GenAI: Contaction

Most relevant contributions of the founder

Decades before everyone was talking about “Human-Centered AI” and using AI-based apps every day, Anthony Jameson was conducting and publishing in-depth research on how to design interactions between human and artificial intelligence — against a background of education in philosophy (Harvard) and psychology (Universities of Hamburg and Amsterdam) and while working in internationally renowned AI projects.

Google Scholar profile page
Archive version of the former DFKI homepage

The following contributions are especially relevant to Contaction's work:

1. Fundamental early contribution to the pragmatics of natural language dialog systems

Invention and realization of the anticipation feedback loop: A system generating text aims to avoid misunderstanding by anticipating how its text will be interpreted by readers/listeners, taking into account their likely assumptions about the system’s bias.

This research is now more relevant than ever with the advent of AI models that are immeasurably more capable at text generation and understanding than those of the early 1980s but which come with a danger of being misinterpreted.

Jameson, A., & Wahlster, W. (1982). User modelling in anaphora generation: Ellipsis and definite description. Proceedings of the First European Conference on Artificial Intelligence, Orsay, France, pp. 222-227.

Jameson, A. (1989). But what will the listener think? Belief ascription and image maintenance in dialog. In A. Kobsa & W. Wahlster (Eds.), User models in dialog systems (pp. 255-312). Berlin: Springer.

2. Influential models of usability issues raised by interaction with AI

Publication of widely cited analyses of the possible negative side-effects of systems that adapt to their users—and of interactive intelligent systems more generally.

These analyses systematically illuminate the potentially problematic phenomena in interaction with AI-based systems: their typical causes and how they can be prevented; their typical consequences and how they can be mitigated; and the evolution of these phenomena over time.

Jameson, A. (2008). Adaptive interfaces and agents. In A. Sears & J. A. Jacko (Eds.), The human-computer interaction handbook: Fundamentals, evolving technologies and emerging applications (2nd edition) (pp. 433-458). Boca Raton, FL: CRC Press.

Jameson, A. (2009). Understanding and dealing with usability side effects of intelligent processing. AI Magazine, 30(4), 23-40.

3. Widely adopted model of the human-related challenges in recommendation to groups

Publication in 2007 of of the most widely cited comprehensive analysis to date of the problem of generating and presenting recommendations to groups.

The analysis is still widely used today, because of its forward-looking focus on human factors as well as technical issues.

Jameson, A., & Smyth, B. (2007). Recommendation to groups. In P. Brusilovsky, A. Kobsa, & W. Nejdl (Eds.), The adaptive web: Methods and strategies of web personalization (pp. 596-627). Berlin: Springer.

4. Founding in 2009 (with John Riedl) of the ACM Transactions on Interactive Intelligent Systems (TiiS)

ACM TiiS is the only top-tier journal devoted entirely to interactions between humans and AI-based systems.

The six years as founding coeditor-in-chief involved collaboration with hundreds of relevant researchers around the world.

Current web pages for ACM TiiS

5. Comprehensive synthesis for the HCI field of scientific knowledge about human choice and choice support

Publication in 2014 of the only book-length comprehensive synthesis for designers of interactive systems of knowledge from psychology and other fields about how people make choices and how they can be helped to make better choices.

Because of the complexity of the topic and the vast and diverse amount of relevant research that exists, most people who design and evaluate interactive systems have difficulty getting a comprehensive and representative understanding of the relevant issues.

This knowledge is especially relevant now that may people agree that AI should aim to help people make better choices themselves instead of just making choices on their behalf.

Jameson, A., Berendt, B., Gabrielli, S., Gena, C., Cena, F., Vernero, F., & Reinecke, K. (2014). Choice architecture for human-computer interaction. Foundations and Trends in Human-Computer Interaction, 7(1-2), 1-235.

6. Introduction of an interaction paradigm for supporting human trial and error in large search spaces

Invention of “Parallel Exploration” and deployment in multinational projects.

This paradigm enables individuals and groups to systematically explore multiple options in parallel.

It was a key part of the large-scale research project, led by Jameson, that won the 2015 Semantic Web Challenge.

Troncy, R., Rizzo, G., Jameson, A., et al. (2017). 3cixty: Building comprehensive knowledge bases for city exploration. Journal of Web Semantics, 46-47, 2-13.

‍

7. Unique integration of research on individual and group decision making and recommender systems

Publication in 2022 of the only comprehensive analysis to date of the relationships between individual and group decision making and recommender systems.

This analysis generalizes the models in the book Choice architecture for human-computer interaction to cover group decision making and analyses the many different ways in which recommender systems can support individual and group decisions.

Jameson, A., Willemsen, M., & Felfernig, A. (2022). Individual and group decision making and recommender systems. In F. Ricci, L. Rokach, & B. Shapira (Eds.), Recommender systems handbook (3nd edition) (p. 789-832). Berlin: Springer.

A spin-off of Germany's leading AI research center ...

... making (AI-generated) content easily available to collaborating groups

Team

Most relevant contributions of the founder