Back to all keywords
CHI
21 items tagged with “CHI”
Essays
- CHI 2026 | Add an AI Classmate — Small Classes of Two May Benefit MostThis CHI 2026 paper designs AI as a teacher-triggered virtual peer and observes how it reshapes student talk in Japanese elementary classrooms.
- CHI 2026 | Emotion Annotation Need Not Label Everything — Catch Experience Turning PointsPREFAB uses preference learning to find emotion turning points, so self-annotation can label less while aiming to keep annotation quality.
- CHI 2026 | Dashboard Too Complex? Don't Only Add a Chat Box — Let AI PointDiana combines voice, text, mouse pointing, and visual highlighting so users can onboard to a dashboard the way they would ask an expert.
- CHI 2026 | High Schoolers Find Scratch Childish — What Blocks Them Is a Sense of GrowthHigh schoolers dislike block-based programming not because it is too simple, but because the curriculum never lets them see themselves growing.
- CHI 2026 | Helping Kids Reflect with AI Is Hardest Not in Follow-ups, but in Knowing When to StopRules keep the teaching path on track; the LLM handles flexible follow-ups. A robotics camp study finds that responding more does not mean understanding deeper.
- CHI 2026 | Everyone Has an LLM — Why Do Technical People Still Win?Even with the same LLM, technical experience still strongly predicts performance. What really widens the gap is planning, expression, understanding, and verification.
- CHI 2026 | Chat with Ocean Animals for Three Turns — Do People Become Greener? Only Half RightA 900-person randomized experiment finds AI ocean characters can boost pro-environmental intention and green choices, yet struggle to shift deeper climate attitudes.
- CHI 2026 | An AI Persona Is Not Behavior Control — Bias Still Needs CalibrationCoBRA turns classic social experiments into a behavioral ruler for agents, making cognitive bias a measurable, tunable, reproducible control variable instead of a prompt wish.
- CHI 2026 | AI Design Tools Should Leave a Trail of Team Thinking, Not Only ResultsCoNode organizes AI generation into trackable, reusable, recombinable team knowledge, and validates its value with two rounds of user research.
- CHI 2026 | AI Can Save Teachers Time — But Not Classroom JudgmentInterviews with 22 teachers show that generative AI can ease repetitive planning work, yet what teachers guard most carefully is professional judgment, relationships, and responsibility.
- CHI 2024 | If AI Knows Less, Do Students Explain More Deeply?When students teach AI in turn, an overly capable model steals explaining opportunities. TeachYou uses controlled knowledge states, active probing, and teaching feedback to make knowledge-building dialogue denser.
- CHI 2026 | Why Literature Reviews Get Messier — The Switch Between Figures and TextCrossLit’s user study suggests visual layout suits growing structure, while text suits polishing argument; the real design problem is when to sync the two — and when to stop.
- CHI 2026 | Showing AI Debate May Not Build Trust — Transparency Needs to Be Just RightA 12-person qualitative study shows multi-agent consensus, disagreement, and critique are all double-edged. Users do not want to see all thinking — they want enough information for the moment.
- CHI 2026 | When AI Speaks in More Detail, Why Do Users Also Say More?A picture-matching collaboration experiment with 267 people found that when agents give fuller cues, people spend less effort listening to instructions — yet when they speak, they also give more detail. What matters is not only how much is said, but what is said and what both sides can see.
- CHI 2026 | AI Helps Teachers Write Feedback — Drafts Improve, But Do Students Learn?In a classroom trial with 354 students, after TAs reviewed AI suggestions, revision quality rose, but the post-test found no significant between-group difference, and grading was not faster. Which step did the comments concretely help?
- CHI 2026 | To Make a Wall Obey, Why Must It Be Allowed to Guess What You Mean?How does a shape-changing wall understand “a bit lower”? A design workshop with 20 people and a Wizard-of-Oz study with 12 expose the tension between control and unspoken intent.
- CHI 2026 | Without Blocking Friends, Can AI Soften Only the Details That Hurt?Want to keep friends’ posts while avoiding details that hurt. DIY-MOD tries transforming content on the receiver side; two rounds of user research also exposed inconsistent preferences, changed original meaning, and generation failures.
- CHI 2026 | Kids Can Build the Game — But May Not Know How to Change ItAI let children build features they had not learned, yet did not necessarily help them understand or revise code. A Scratch programming study with children records the concrete path from creative excitement to getting stuck while debugging.
- CHI 2026 | Coding Agents Save Effort — Why Most Participants Still Prefer CopilotIn a controlled study of 20 participants, the Agent condition yielded higher task correctness and less estimated human effort time, yet future-use preference still leaned toward Copilot. After automation executes, understanding and control still need a person in charge.
- CHI 2026 | Asking AI for Inspiration — Why Waiting for Answers Is Not EnoughA CHI 2026 survey of 85 studies extends design-by-analogy from inspiration into prototyping, fabrication, and evaluation. What matters is what you borrow, how you borrow it, and where users can still redirect the process.
- CHI 2026 | AI Sparked More Ideas — Why Did 18 of 30 Students Converge on Similar Designs?The AI group left 291 ideas versus 89 for sketching, yet image concepts still converged. This classroom study of 61 design students separates idea count, design intent, and creativity ratings.