CHI 2026HONORABLE MENTION
Bhada Yun, Renn Su, April Yi Wang
Does AI understand human values? While this remains an open philosophical question, we take a pragmatic stance by introducing VAPT, the Value-Alignment Perception Toolkit, for studying how LLMs reflect people's values and how people judge those reflections. 20 participants texted a chatbot over a month, then completed a 2-hour interview with our toolkit evaluating AI's ability to extract (pull details regarding), embody (make decisions guided by), and explain (provide proof of) their values. 13 participants ultimately left our study convinced that AI can understand human values. Thus, we warn about "weaponized empathy": a design pattern that may arise in interactions with value-aware, yet welfare-misaligned conversational agents. VAPT offers a new way to evaluate value-alignment in AI systems. We also offer design implications to evaluate and responsibly build AI systems with transparency and safeguards as AI capabilities grow more inscrutable, ubiquitous, and posthuman into the future.
Read paper summaryarrow_forwardCHI 2026
Jazmin Collins, Sharon Y Lin, Tianqi Liu, Andrea Stevenson Won
As social virtual reality (VR) grows more popular, addressing accessibility for blind and low vision (BLV) users is increasingly critical. Researchers have proposed an AI “sighted guide” to help users navigate VR and answer their questions, but it has not been studied with users. To address this gap, we developed a large language model (LLM)-powered guide and studied its use with 16 BLV participants in virtual environments with confederates posing as other users. We found that when alone, participants treated the guide as a tool, but treated it companionably around others, giving it nicknames, rationalizing its mistakes with its appearance, and encouraging confederate-guide interaction. Our work furthers understanding of guides as a versatile method for VR accessibility and presents design recommendations for future guides.
Read paper summaryarrow_forwardCHI 2026
Shreyan Biswas, Alexander Erlei, Ujwal Gadiraju
Large language models (LLMs) increasingly support heterogeneous tasks within a single interface, requiring users to form, update, and act upon beliefs about one system across domains with different reliability profiles. Understanding how such beliefs transfer across tasks and shape delegation is critical for the design of multipurpose AI systems. We report a preregistered experiment (N = 240, 7,200 trials) in which participants interacted with a controlled AI simulation across grammar checking, travel planning, and visual question answering. Delegation was operationalized as a binary reliance decision—accepting the AI’s output versus acting independently—and belief dynamics were evaluated against Bayesian benchmarks. We find three main results. First, participants do not reset beliefs between tasks, instead carrying expectations from prior interactions. Second, within tasks, belief updating follows the Bayesian direction but is substantially conservative. Third, delegation is driven primarily by subjective beliefs about AI accuracy rather than self-confidence, though confidence independently reduces reliance when beliefs are held constant. Based on these results, we discuss implications for expectation calibration, reliance design, and the risks of belief spillovers in deployed LLM-based interfaces.
Read paper summaryarrow_forwardCHI 2026
Xueqing Li, Danqi Huang, Tianyu Yu, Shuzi Yin
We introduce DuoMorph, a design and fabrication method that synergistically integrates Fused Deposition Modeling(FDM) printing and pneumatic actuation to create novel shape-changing interfaces. In DuoMorph, the printed structures and heat-sealed pneumatic elements are mutually designed to actuate and constrain each other, enabling functions that are difficult for either component to achieve in isolation. Moreover, the entire hybrid structure can be fabricated through a single, seamless process using only a standard FDM printer—including both heat-sealing and 3D/4D printing. In this paper, we define a design space including four primitive categories that capture the fundamental ways in which printed and pneumatic components can interact. To support this process, we present a fabrication method and an accompanying design tool. Finally, we demonstrate the potential of DuoMorph through example applications and performance demonstrations.
Read paper summaryarrow_forwardCHI 2026
Qi Zhang, Shuwen Jiang, Zeshui Li, Yong Lyu
Creating interactive papercrafts often involves the processes of craft making (e.g., folding, cutting, gluing, etc.) and fabricating the embedded functional circuitry. These two processes are usually separated in the current practice, making the workflow laborious and affecting the in-paper circuit stability. To address the issue of the separated crafting and circuiting processes, we present LiqMetCraft, a toolkit for creating electronics-embedded papercrafts through an integrated process. The toolkit allows users to construct the craft structure by folding and cutting, and forms the circuit traces simultaneously. This is achieved with liquid-metal-dyed paper-like fabric which partially becomes conductive due to the cutting/folding-induced pressure while the unpressed parts of the paper remain insulated. The toolkit consists of software interfaces for papercraft design and hardware components, mainly the liquid-metal-dyed paper-like fabrics and other off-the-shelf components, for physical prototyping. The user studies shows that the participants quickly learned the toolkit and found the integrated process of circuit assembly and shape formation to be engaging and inspiring.
Read paper summaryarrow_forwardCHI 2026
Jian Zhang, Wafa Johal, Jarrod Knibbe
Tangible interactions involve multiple sensory cues, enabling the accurate perception of object properties, such as size. Research has shown, however, that if we decouple these cues (for example, by altering the visual cue), then the resulting discrepancies present new opportunities for interactions. Perception over time though, not only relies on momentary sensory cues, but also on a priori beliefs about the object, implying a continuing update cycle. This cycle is poorly understood and its impact on interaction remains unknown. We study (N=80) visuo-haptic perception of size over time and (a) reveal how perception drifts, (b) examine the effects of visual priming and dead-reckoning, and (c) present a model of visuo-haptic perception as a cyclical, self-adjusting system. Our work has a direct impact on illusory perception in VR, but also sheds light on how our visual and haptic systems cooperate and diverge.
Read paper summaryarrow_forwardCHI 2026
Menglin Zhao, Zhuorui Yong, Ruijia Guan, Kai-Wei Chang
Serious Illness Conversations (SICs)—discussions about values and care preferences for patients with life-threatening illness—rarely occur in Emergency Departments (EDs), despite evidence that early conversations improve care alignment and reduce unnecessary interventions. We interviewed 11 ED providers to identify challenges in SICs and opportunities for technology support, with a focus on AI. Our analysis revealed a four-stage SIC workflow (identification, preparation, conduction, documentation) and barriers at each stage, including fragmented patient information, limited time and space, lack of conversational guidance, and burdensome documentation. Providers expressed interest in AI systems for synthesizing information, supporting real-time conversations, and automating documentation, but emphasized concerns about preserving human connection and clinical autonomy. This tension highlights the need for technologies that enhance efficiency without undermining the interpersonal nature of SICs. We propose design guidelines for ambient and peripheral AI systems to support providers while preserving the essential humanity of these conversations.
Read paper summaryarrow_forwardCHI 2026
Xiaohan Huang, Jiahui Wu, Ming Zhou, Xuemin Zhang
Error awareness—the ability to detect errors, adjust strategies, and prevent mistakes—is critical in high-stakes human–computer interaction (e.g., aviation, autonomous system supervision), as well as in everyday life. However, this ability deteriorates under heavy cognitive load, and effective countermeasures remain scarce. We investigate whether transcranial direct current stimulation (tDCS) can mitigate this deficit. Using a multi-rule task with EEG, we found that under high load, tDCS over the left dorsolateral prefrontal cortex (DLPFC) significantly improved error awareness, reflected in both behavioral measures and a neural index. Crucially, mediation analysis showed this effect was achieved by improving working memory capacity, facilitating better real-time error detection. Our findings demonstrate that neurostimulation sustains self-monitoring by augmenting depleted cognitive resources. We formalize this in the Dynamic Cognitive Resource Barrel Theory: error awareness is limited by the most depleted cognitive “stave” after primary task demands. These results offer a principled path for designing neuroadaptive systems that predict and support these processes in critical moments.
Read paper summaryarrow_forwardCHI 2026HONORABLE MENTION
Nishchal Jagadeesha, Chorong Park, Avery Kruppe, Yanfu Liu
Social robots in the home bring new privacy risks and concerns for older adults. Yet, current technology privacy mechanisms typically use a one-time and universal consent mechanism (e.g., user agreement checkbox, browser cookie setting, etc.), lacking consideration of how privacy is holistically experienced. Designing for privacy requires a multidimensional approach to support how older adults experience privacy. To investigate older adult-centered privacy mechanisms for social robots, we conducted two participatory design (PD) workshops at local assisted living facilities. Our findings from these workshops suggest that older adults do not treat privacy as static, but as a temporal and situational practice that requires continuous negotiations and revisions. We subsequently conducted a post-PD speculative design (SD) process that extracted three design features for privacy—aware social robots-privacy profiles, real-time privacy feedback, and data ownership tools—that can support older adults’ multidimensional privacy experiences.
Read paper summaryarrow_forwardCHI 2026
Carla Coutant, Adrien Chaffangeon Caillet, Renaud Blanch
On a visualization, the position of the marks encoding data is the most expressive and effective visual channel. It conveys order and quantity without impairing the perception of other visual channels. In the field of Information Visualization, position is often restricted to two dimensions, because using the third dimension, \emph{depth}, usually affects the perception of size, which is also one of the most effective visual channels. We propose a new visual channel, D-MO (Depth from Motion and Occlusion), a combination of visual cues, \emph{motion} and \emph{occlusion}, with interactions, that induces a depth perception suitable for combined use with classic visual channels. We characterize the expressiveness and effectiveness of D-MO and show that it is a magnitude channel with good accuracy, acceptable discriminability, and is separable from size. Thus, D-MO opens up new areas for visualization design, which is limited by the scarceness of available visual channels.
Read paper summaryarrow_forwardCHI 2026
Jiwan Kim, Chi-Jung Lee, Hohurn Jung, Tianhong Catherine Yu
Tracking hand poses on wrist-wearables enables rich, expressive interactions, yet remains unavailable on commercial smartwatches, as prior implementations rely on external sensors or custom hardware, limiting their real-world applicability. To address this, we present WatchHand, the first continuous 3D hand pose tracking system implemented on off-the-shelf smartwatches using only their built-in speaker and microphone. WatchHand emits inaudible frequency-modulated continuous waves and captures their reflections from the hand. These acoustic signals are processed by a deep-learning model that estimates 3D hand poses for 20 finger joints. We evaluate WatchHand across diverse real-world conditions---multiple smartwatch models, wearing-hands, body postures, noise conditions, pose-variation protocols---and achieve a mean per-joint position error of 7.87 mm in cross-session tests with device remounting. Although performance drops for unseen users or gestures, the model adapts effectively with lightweight fine-tuning on small amounts of data. Overall, WatchHand lowers the barrier to smartwatch-based hand tracking by eliminating additional hardware while enabling robust, always-available interactions on millions of existing devices.
Read paper summaryarrow_forwardCHI 2026
Gyeongdeok Kim, Chungman Lim, Gyungmin Jin, Gunhyuk Park
Independent and secret voting is a constitutional right, yet blind and low-vision voters (BLVs) continue to face barriers when casting their votes. Existing methods such as tactile templates often require braille literacy or assistance, while electronic ballot-marking devices raise cost and security concerns. We present I-VAMOS, a voting assistance system that enables BLVs to cast paper ballots securely and independently. Based on participatory sessions with BLVs, I-VAMOS integrates a ballot slide frame, a spring-loaded stamp, and real-time OCR-based speech and visual feedback, operating offline without the need for customized templates. With the improved I-VAMOS, we conducted a user study (n=16), balanced across vision status, braille literacy, and age. Results showed that I-VAMOS significantly reduced workload (NASA–TLX; 26.1) and improved stamping accuracy (91.7%) and usability (SUS; 79.1) compared to existing aids, though took longer completion times (29.6s). These findings emphasize that I-VAMOS enables independent and confidential voting for BLVs.
Read paper summaryarrow_forward