pHAI@HCOMP2023: Prototyping Human-AI Interactions

This half-day tutorial/workshop aims to bring together designers, engineers, and practitioners to explore ways to support the prototyping of interactions between human users and AI/ML models. Through a combination of paper and screen-based tools, we will present our modeling approach for designing Human-AI interactions and guide participants in creating their own representations and interactions. The goal of this workshop is to raise the discussion and get insights towrads bridging the gap between engineering and design practices while designing human-AI interactions -- Link to tool

This site used HTML5 Templates.

Workshop Description

Workshop participants will be actively involved in group activities, including design exercises using card-based and screen-based materials, round-table discussions, and feedback sessions. Participants will follow structured paper-and-card based design exercises that articulate the core concepts of HAI interactions, and then adapt these to their own case studies. They will also use a screen-based tool that supports model-informed prototyping, to work towards higher fidelity understandings of interactions with models. Based on the activities, we will discuss the implementation and design challenges that may arise when prototyping HAI interactions.

The main outcome of the workshop will be the interactions with the participants considering current trends and challenges towards the development of HAI interactions. We will produce a set of example HAI interaction designs based on participant’s case studies that can be brought together as a collaborative publication. Our goal is to develop insights towards applying best practices based on existing frameworks and methodologies, including Responsible and Trustworthy AI.

List or materials (guidelines/frameworks for HAI interactions):

Participation - Submit your own Materials

Workshop participants can submit their own use case to be included in the workshop activities. In this way, participants will have the opportunity to design interactions and get insights from each other for a variety of use cases – including their own. Participants will have the chance to present their own use cases and scenarios, and receive feedback, through the group activities and discussions. We expect that such activities will provide us with insights about new ways to prototype interactions between users and models.

The use case description should be a 1-2 pages PDF file (see example) and include the following:

Participant information: name(s), affiliation(s) and email adresse(s) of participant(s)
Overview: a short description of the context in which models are to be deployed, and the goals of the interaction.
Description of the AI/ML model(s): what model(s) form the core of your case study? What are their inputs and outputs, explanations or interactivity do they provide, and what is important to know about their internal functioning?
Description of the human user(s): what are the various roles present, both directly and indirectly affected by the models?
Scenarios for interactions between user(s) and model(s): what interaction scenarios do you initially have in mind?
Optional: figures/links for model, data and interactions

Based on the submitted use cases, our goal is to select and define a variety of models and application contexts to be included as workshop activities.

Note: Workshop participation is not dependent on having an accepted case study. Conference participants will be welcome to register and participate to the workshop.

Deadline for submission of use case: October 4, 2023

Notification date: October 5, 2023

Submit your Use Case: Submission Form

pHAI@HCOMP2023: Prototyping Human-AI Interactions

Workshop Description

Schedule (Tentantive)

Participation - Submit your own Materials

Contact