Half-Day Workshop (in the afternoon)
(Exact time will depend on IDC scheduling, but here is the structure in their format.)

Session 1 — Framing & shared ground (60 mins)

13:00 – 13:15 (15 mins) — Welcome & Framing

Brief overview of workshop aims, goals, and schedule. Introduction of organisers and themes: trust, design, and evaluation in child–AI interaction.

Aims of the workshop

Why trust evaluation in child-AI matters

What participants will produce today

Quick overview of the dataset

13:15 – 13:35 (20 mins) — Interactive Provocation by Jason Yip

Short stimulus talk + discussion:

How do children express trust?

Is compliance the same as trust?

What developmental differences matter?

When does AI literacy mask over-trust?

Open floor reactions (8-10 mins)

13:35 – 14:00 (25 mins) — Rapid Transcript Annotation Sprint 

Move quickly from observation - interpretation - metric potential

Step 1 (10 mins) - Individual Marking on short transcript excerpts

Step 2 (10 mins) - Small group clustering

Step 3 (5 mins) - Metric seeds

14:00 – 14:30 (30 mins) — Coffee Break

14:30 – 15:10 (40 mins) — Transcript Analysis (Small Groups)

Each group produces a visual map of 'Trust Signals'

15:10 – 15:45 (35 mins) — Co-Design Evaluation Sprint 

Groups prototype developmentally sensitive evaluation artefacts, such as:

• markers of trust and distrust

• transparency and explainability cues

• child-friendly evaluation prompts

• cross-age considerations (6–15)

15:45 – 16:00 (15 mins) — Gallery Walk

Groups rotate, observe others' artefacts and reflect on

• over-simplifying children’s trust

• bias toward certain age groups

• ignoring cultural/language differences

• misinterpreting emotional cues

Groups refine ideas based on identified risks.

16:00– 16:30 (30 mins) — Reverse Brainstorming

“How might our evaluation approach fail children?”

16:30– 16:50 (20 mins) — Collective Synthesis

Plenary discussion

16:50– 17:00 (10 mins) — Closing & Next Steps

Open repository

Follow-up online session