How iSchool Students Picture GenAI

OVERVIEW

This published study examined nine participants’ drawings and explanations to investigate their mental models of GenAI. I led the follow-on investigation outside the class, including thematic analysis; the paper is coauthored by Irene Lopatovska, Conor Mack, and Ellen Connors.

Hand-drawn sketch titled 'How GenAI works,' showing input, processing, and output stages connected by arrows

TIMELINE

Published 2026

ROLE

Researcher & Co-author

TEAM

Irene Lopatovska, Conor Mack, Ellen Connors

FOCUS

UX Research

Drawings and explanations

Drawings from the published nine-participant qualitative study.

The study examined drawings and explanations from nine iSchool participants. I led this follow-on investigation outside the class, analyzing the material through thematic coding. The paper retains its three coauthor credits.

The related class-based study behind The normalization of unease involved my research-design and data-collection contribution as a teaching assistant. The two publications share broader research context and should not be read as entirely separate recruitment efforts.

What the analysis suggested

Participant D8's drawing depicting generative AI as a masked figure stealing art, writing, and research into a pile, with a separate panel showing the AI training overnight

Participant D8's drawing: GenAI as theft, not generation.

Participants drew on familiar information-retrieval models and human analogies to explain GenAI. The analysis found limited visibility into the people involved in producing these systems.

These findings concern the explanations examined in this qualitative study. They do not establish how common a mental model is among all students or GenAI users, or a causal link between understanding and trust.

Qualitative themes, not validated segments

The drawings and explanations offered different ways of picturing GenAI, including familiar search-system models and human analogies. The analysis also identified limited visibility into the human involvement behind these systems.

These qualitative themes can inform questions for further research. They are not validated user segments or population estimates.

Implications for AI literacy and explanations

The findings suggest a role for AI literacy and interfaces that make system behavior and human involvement easier to understand. Familiar search-system analogies can offer a starting point for explanations while also obscuring how generation works.

These are implications for further design and research, not tested interventions or demonstrated improvements in comprehension, trust, or adoption.

Publication and limits

Lopatovska, I., Mack, C., and Connors, E. (2026). iSchool students’ mental models of GenAI. Information Research, 31(ISIC), 338–355. DOI: 10.47989/ir31ISIC65224. The publication record is established; this account does not assign a conference presentation to a particular coauthor.

Nine participants and shared broader study context limit generalization. Further research would be needed to assess whether explanations change understanding or comfort with GenAI.

This page is looked after by the Archivist, an agent I built that works from my primary research and can’t publish without me. Here’s how that works.

How iSchool Students Picture GenAI

OVERVIEW

This published study examined nine participants’ drawings and explanations to investigate their mental models of GenAI. I led the follow-on investigation outside the class, including thematic analysis; the paper is coauthored by Irene Lopatovska, Conor Mack, and Ellen Connors.

Hand-drawn sketch titled 'How GenAI works,' showing input, processing, and output stages connected by arrows

TIMELINE

Published 2026

ROLE

Researcher & Co-author

TEAM

Irene Lopatovska, Conor Mack, Ellen Connors

FOCUS

UX Research

Drawings and explanations

Drawings from the published nine-participant qualitative study.

The study examined drawings and explanations from nine iSchool participants. I led this follow-on investigation outside the class, analyzing the material through thematic coding. The paper retains its three coauthor credits.

The related class-based study behind The normalization of unease involved my research-design and data-collection contribution as a teaching assistant. The two publications share broader research context and should not be read as entirely separate recruitment efforts.

What the analysis suggested

Participant D8's drawing depicting generative AI as a masked figure stealing art, writing, and research into a pile, with a separate panel showing the AI training overnight

Participant D8's drawing: GenAI as theft, not generation.

Participants drew on familiar information-retrieval models and human analogies to explain GenAI. The analysis found limited visibility into the people involved in producing these systems.

These findings concern the explanations examined in this qualitative study. They do not establish how common a mental model is among all students or GenAI users, or a causal link between understanding and trust.

Qualitative themes, not validated segments

The drawings and explanations offered different ways of picturing GenAI, including familiar search-system models and human analogies. The analysis also identified limited visibility into the human involvement behind these systems.

These qualitative themes can inform questions for further research. They are not validated user segments or population estimates.

Implications for AI literacy and explanations

The findings suggest a role for AI literacy and interfaces that make system behavior and human involvement easier to understand. Familiar search-system analogies can offer a starting point for explanations while also obscuring how generation works.

These are implications for further design and research, not tested interventions or demonstrated improvements in comprehension, trust, or adoption.

Publication and limits

Lopatovska, I., Mack, C., and Connors, E. (2026). iSchool students’ mental models of GenAI. Information Research, 31(ISIC), 338–355. DOI: 10.47989/ir31ISIC65224. The publication record is established; this account does not assign a conference presentation to a particular coauthor.

Nine participants and shared broader study context limit generalization. Further research would be needed to assess whether explanations change understanding or comfort with GenAI.

This page is looked after by the Archivist, an agent I built that works from my primary research and can’t publish without me. Here’s how that works.

How iSchool Students Picture GenAI

OVERVIEW

This published study examined nine participants’ drawings and explanations to investigate their mental models of GenAI. I led the follow-on investigation outside the class, including thematic analysis; the paper is coauthored by Irene Lopatovska, Conor Mack, and Ellen Connors.

Hand-drawn sketch titled 'How GenAI works,' showing input, processing, and output stages connected by arrows

TIMELINE

Published 2026

ROLE

Researcher & Co-author

TEAM

Irene Lopatovska, Conor Mack, Ellen Connors

FOCUS

UX Research

Drawings and explanations

Drawings from the published nine-participant qualitative study.

The study examined drawings and explanations from nine iSchool participants. I led this follow-on investigation outside the class, analyzing the material through thematic coding. The paper retains its three coauthor credits.

The related class-based study behind The normalization of unease involved my research-design and data-collection contribution as a teaching assistant. The two publications share broader research context and should not be read as entirely separate recruitment efforts.

What the analysis suggested

Participant D8's drawing depicting generative AI as a masked figure stealing art, writing, and research into a pile, with a separate panel showing the AI training overnight

Participant D8's drawing: GenAI as theft, not generation.

Participants drew on familiar information-retrieval models and human analogies to explain GenAI. The analysis found limited visibility into the people involved in producing these systems.

These findings concern the explanations examined in this qualitative study. They do not establish how common a mental model is among all students or GenAI users, or a causal link between understanding and trust.

Qualitative themes, not validated segments

The drawings and explanations offered different ways of picturing GenAI, including familiar search-system models and human analogies. The analysis also identified limited visibility into the human involvement behind these systems.

These qualitative themes can inform questions for further research. They are not validated user segments or population estimates.

Implications for AI literacy and explanations

The findings suggest a role for AI literacy and interfaces that make system behavior and human involvement easier to understand. Familiar search-system analogies can offer a starting point for explanations while also obscuring how generation works.

These are implications for further design and research, not tested interventions or demonstrated improvements in comprehension, trust, or adoption.

Publication and limits

Lopatovska, I., Mack, C., and Connors, E. (2026). iSchool students’ mental models of GenAI. Information Research, 31(ISIC), 338–355. DOI: 10.47989/ir31ISIC65224. The publication record is established; this account does not assign a conference presentation to a particular coauthor.

Nine participants and shared broader study context limit generalization. Further research would be needed to assess whether explanations change understanding or comfort with GenAI.

This page is looked after by the Archivist, an agent I built that works from my primary research and can’t publish without me. Here’s how that works.