MM/AIWorkshop

Mental Models of AI

Workshop on Mental Models in Human-AI Interaction: Methods and Challenges in the Generative and Agentic AI Era

Where
🇫🇮 Helsinki, Finland
When
February 8, 2027
Co-located with
ACM IUI 2027 ↗

Important Dates

Important DatesAll deadlines AoE
  1. 01Paper submission deadline· upcomingNov 9, 2026
  2. 02Notification of acceptanceDec 7, 2026
  3. 03Camera-ready paper deadlineDec 21, 2026
  4. 04Workshop at ACM IUI 2027, HelsinkiFeb 8, 2027

About this workshop

The mental model construct is widely used in HCI to refer to the knowledge structure people hold to reason about computing systems. Yet the construct is often used interchangeably with related concepts (e.g. folk theories, sensemaking) and methods of studying it are many and diverse. Generative and agentic AI systems further complicate mental model formation and elicitation as such systems are opaque by design and increasingly act on users' behalf across files, applications, and on the web. Together, these challenges may hinder the commensurability of research on people’s mental models of AI systems. The MM/AI workshop calls for a critical reassessment of how we understand and study mental models in human-AI interaction research. It aims to foster theoretical and methodological exchange on mental models in human–AI interaction, identify open challenges, and develop directions for future research.

Read the full proposal ↗

Call for papers

How should we understand and study people's mental models of AI?

We invite researchers and practitioners to the inaugural Mental Models of AI (MM/AI) workshop, held on 8 February 2027 at ACM IUI 2027 in Helsinki, Finland.

We welcome theoretical, empirical, methodological, and conceptual contributions on mental models in human–AI interaction. We are particularly interested in empirical work that critically reflects on how mental models are conceptualized and elicited, and on the assumptions and limitations of the methods used to study them.

Page limit
4 pages, excluding referencesReferences do not count toward the limit.
Template
Latest ACM two-column template, anonymized · template ↗Strip author names, affiliations, and acknowledgements before submitting.
Review
Each paper will receive two double-blind peer reviews on OpenReviewReviews prioritize relevance to the workshop themes, methodological and theoretical insight, and potential to stimulate discussion. Submissions may build on previously published work, provided they are identified as such in the final, deanonymized version and the submission offers new reflection on the mental model construct, its elicitation, or the implications of the original findings.
Submission portal
OpenReview ↗
Questions
workshop@mentalmodelsofai.com

Workshop Program

Length

Half-day (3.5 hours), in person at IUI '27.

Capacity

25–30 participants: authors of accepted submissions, co-organizers, and attendees admitted through open registration.

Indicative schedule
  1. 0:00 – 0:15Workshop presentation and a short genealogy of the mental model construct by Dr. Téo Sanchez
  2. 0:15 – 0:50Selected lightning talks: five submissions presented in a 5-minute followed by 2 minutes of questions
  3. 0:50 – 1:30Peer elicitation exercise: participants work in pairs to elicit each other's mental model of a coding agent or a text-to-image generator.
  4. 1:30 – 2:00Coffee break (the previous activity may overlap if needed)
  5. 2:00 – 3:00Breakout discussions: groups tackle open questions on the future of mental model research in the context of contemporary AI systems
  6. 3:00 – 3:30Report-back and closing discussion: each group shares insights, points of disagreement, and priorities for future research

After acceptance

  • 01

    The workshop is non-archival

    Accepted papers are not published in formal proceedings, so presenting here does not preclude submitting the work elsewhere.

  • 02

    Archiving is opt-in

    With the authors' consent, accepted papers are listed on this website. Authors who would rather self-archive on Zenodo, arXiv, or a repository of their choice are free to do so.

  • 03

    Camera-ready by December 21, 2026

    Deanonymize the paper and add author names, affiliations, and acknowledgements. Work that builds on a previous publication should identify it as such in this version.

  • 04

    Present in person in Helsinki

    The day is built around discussion and hands-on exercises, so we expect at least one author of each accepted paper on site. Five submissions are selected for a five-minute lightning talk; all participants take part in the peer elicitation exercise and the breakout discussions.

Organizers & Program Committee

The organizing team combines first-hand research experience about mental models in human-AI interaction and a strong motivation to reopen discussions within the IUI community. Click any organizer for their full bio and profile links.

Program Committee