Who bears the cost of honesty?
A CRAFT workshop on when AI disclosure builds trust, when it creates stigma, and how transparency policies can be designed more fairly.Jun 26, 3:00 - 5:30 PM · Le Centre Sheraton Montréal
Disclosure is not a neutral act.
AI disclosure is increasingly being written into policy. That can support transparency, but evidence also shows that revealing AI use can stigmatize people and their work, with heavier costs for minoritized groups. This workshop asks how transparency can be designed without shifting those costs onto the people asked to disclose.
- 01
Accountability still matters.
AI disclosure can clarify how work was made, support accountability, and help people decide when to trust it.
- 02
But disclosure is not neutral.
The same disclosure can also trigger suspicion: language support read as cheating, accommodation read as lower competence, or early ideation read as automated authorship.
- 03
So disclosure needs better design.
Fair policies should consider context, power, and privacy, especially for people already more likely to face stigma.
We move beyond the question “should people disclose AI use?”
Together, we ask:
- Who benefits from disclosure, and who carries its costs?
- Which needs, rights, and risks are in tension?
- How can disclosure policies provide useful clarity without unnecessary exposure?
This is where disclosure gets complicated.
Six everyday situations. Four ways to read each one. Choose a lens, and the same scenario reveals a different need, risk, or design opportunity.
Where social costs can land.
A hands-on workshop, together.
We start with a short framing talk, then move into two participatory activities: Power Mapping and the Disclosure Design Fiction Studio. Groups of 5-6 will work with both digital tools and paper materials.
- 20 min · step 01
Framing the landscape
A local invited researcher introduces disclosure mandates, emerging norms, and evidence of harm.
- 45 min · step 02
Power Mapping
Groups take stakeholder roles, draft future headlines, and map needs, constraints, and power asymmetries.
- 45 min · step 03
Disclosure Design Fiction Studio
Groups sketch or prototype a 2029 disclosure artifact, then complete a Context Card naming benefits, mitigated costs, and remaining harms.
- 10 min · step 04
Share-out
Groups share artifacts and Context Cards; organizers use them as source material for a public equitable-disclosure toolkit.
Ways to identify and reduce disclosure harms
Practical prompts for spotting where disclosure helps some people while increasing risk for others, especially minoritized groups.
Design artifacts and Context Cards
Sketches, prototypes, and notes that connect design choices to accountability, consent, and harm.
A public toolkit
Organizers will synthesize outcomes into a toolkit of considerations, templates, and examples shared back to participants and the public.
The organizing team.
Seven researchers and practitioners across two companies and five academic institutions, working across design, philosophy, history, and AI ethics.
Finola Finn
Postdoctoral ResearcherUniversity of LuxembourgEsch-sur-Alzette, Luxembourgresearch focusAI in historical and creative practice
Angel Hsing-Chi Hwang
Assistant ProfessorUniversity of Southern CaliforniaLos Angeles, USAresearch focusHuman–AI interaction
Donal Khosrowi
Postdoctoral ResearcherLeibniz University Hannover (CELLS)Hanover, Germanyresearch focusEthics of AI and AI-in-science
Seyun Kim
PhD CandidateCarnegie Mellon University, HCIIPittsburgh, USAresearch focusHuman–AI interaction in high-stakes decision-making
Morgan Klaus Scheuerman
Research ScientistSony AIBarcelona, Spainresearch focusGenerative AI in creative industries
Harry Ye
PhD StudentUniversity of TorontoToronto, Canadaresearch focusCritical Engagement with AI, Educational Technology
