Investment decisions are often driven by time pressure and emotion, leaving investors vulnerable to cognitive biases. Mobile trading apps intensify these tendencies, yet existing interventions rely on external constraints that fail to foster lasting behavioral change. We investigate how reflection-centered approaches support mindful decision-making across a spectrum of investor expertise. We present MindStock, a technology probe providing principle-anchored feedback by integrating user-defined principles with behavioral data mirroring trading patterns. In a 6-week field study with 16 investors, we found that meaningful reflection comes from the tension between principles and behavioral data. Principles give context to otherwise opaque metrics, while data keeps principles from drifting into vague self-assurances. This pattern varied by experience: novices gravitated toward normative rule-setting, while experienced investors used the system to test and refine their own assumptions. We contribute design implications for supporting reflection in high-stakes decision-making contexts.

Warning: This article contains stereotypical and offensive content.

Team-based collaboration is a cornerstone of modern creative work. Recent advances in generative AI open possibilities for humans to collaborate with multiple AI agents in distinct roles to address complex creative workflows. Yet, how to form Human-Multi-Agent Teams (HMATs) is underexplored, especially given that inter-agent interactions increase complexity and the risk of unexpected behaviors. In this exploratory study, we aim to understand how to form HMATs for creative work using CrafTeam, a technology probe that allows users to form and collaborate with their teams. We conducted a study with 12 design practitioners, in which participants iterated through a three-step cycle: forming HMATs, ideating with their teams, and reflecting on their teams' ideation. Our findings reveal that while participants initially attempted autonomous team operations, they ultimately adopted team formations in which they directly orchestrated agents. We discuss design considerations for HMAT formation that humans can effectively orchestrate multiple agents.

Large language models (LLMs) are promising tools for scaffolding students’ English writing skills, but their effectiveness in real-time K-12 classrooms remains underexplored. Addressing this gap, our study examines the benefits and limitations of using LLMs as real-time learning support, considering how classroom constraints, such as diverse proficiency levels and limited time, affect their effectiveness. We conducted a deployment study with 157 eighth-grade students in a South Korean middle school English class over six weeks. Our findings reveal that while scaffolding improved students’ ability to compose grammatically correct sentences, this step-by-step approach demotivated lower-proficiency students and increased their system reliance. We also observed challenges to classroom dynamics, where extroverted students often dominated the teacher’s attention, and the system’s assistance made it difficult for teachers to identify struggling students. Based on these findings, we discuss design guidelines for integrating LLMs into real-time writing classes as inclusive educational tools.

AI clones are evolving to include digital representations of real world individuals as chatbots. While often used to replicate famous figures, as the technology becomes more accessible, it is crucial to understand whether everyday users would create their own clones and how they interact with them. In this study, within the scope of AI-generated personas and their role in representing users’ needs and identities, we focus on personas that directly reflect the qualities of real humans. We define this as AI self clones—conversational AI representations that reflect their human creators—and examine how creators construct and engage with them. We conducted a 7-day study in which participants (N=12) created and interacted with their text based AI self clones using CloneBuilder , a web-based authoring interface for configuring and tuning AI self clones. The system enables individuals to create AI representations that encapsulate their unique personality, values, and interaction style. Our findings reveal that each participant developed a clone tailored to their personal circumstances. As the participants iteratively refined and tested their clone, their direction and expectation of AI clones evolved from performing specific roles to evolving entities that facilitated self exploration and relationship formation. Unexpected responses from the clone prompted self reflection and identity questioning. Overall, this paper explores the motivations for creating these clones, the strategies participants use to build and refine them, and the moments of emotional connection and break out experiences that emerge during the crafting process, along with key design implications, challenges, and ethical considerations in developing AI self clones.

Effective feedback, including critique and evaluation, helps designers develop design concepts and refine their ideas, supporting informed decision-making throughout the iterative design process. However, in studio-based design courses, students often struggle to provide feedback due to a lack of confidence and fear of being judged, which limits their ability to develop essential feedback-giving skills. Recent advances in large language models (LLMs) suggest that role-playing with AI agents can let learners engage in multi-turn feedback without the anxiety of external judgment or the time constraints of real-world settings. Yet prior studies have raised concerns that LLMs struggle to behave like real people in role-play scenarios, diminishing the educational benefits of these interactions. Therefore, designing AI-based agents that effectively support learners in practicing and developing intellectual reasoning skills requires more than merely assigning the target persona’s personality and role to the agent. By addressing these issues, we present Feed-O-Meter, a novel system that employs carefully designed LLM-based agents to create an environment in which students can practice giving design feedback. The system enables users to role-play as mentors, providing feedback to an AI mentee and allowing them to reflect on how that feedback impacts the AI mentee’s idea development process. A user study (N=24) indicated that Feed-O-Meter increased participants’ engagement and motivation through role-switching and helped them adjust feedback to be more comprehensible for an AI mentee. Based on these findings, we discuss future directions for designing systems to foster feedback skills in design education.

Hyungyu Shin, Jingyu Tang, Yoonjoo Lee, Nayoung Kim, Hyunseung Lim, Ji Yong Cho, Hwajung Hong, Moontae Lee, Juho Kim. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Patent examination remains an ongoing challenge in the NLP literature even after the advent of large language models (LLMs), as it requires an extensive yet nuanced human judgment on whether a submitted $\textit{claim}$ meets the statutory standards of $\textit{novelty}$ and $\textit{non-obviousness}$ against previously granted claims—$\textit{prior art}$—in expert domains. Previous NLP studies have approached this challenge as a prediction task (e.g., forecasting grant outcomes) with high-level proxies such as similarity metrics or classifiers trained on historical labels. However, this approach often overlooks the step-by-step evaluations that examiners must make with profound information, including rationales for the decisions provided in $\textit{office actions}$ documents, which also makes it harder to measure the current state of techniques in patent review processes. To fill this gap, we construct PANORAMA, a dataset of 8,143 U.S. patent examination records that preserves the full decision trails, including original applications, all cited references, $\textit{Non-Final Rejections}$, and $\textit{Notices of Allowance}$. Also, PANORAMA decomposes the trails into sequential benchmarks that emulate patent professionals' patent review processes and allow researchers to examine large language models' capabilities at each step of them. Our findings indicate that, although LLMs are relatively effective at retrieving relevant prior art and pinpointing the pertinent paragraphs, they struggle to assess the novelty and non-obviousness of patent claims. We discuss these results and argue that advancing NLP, including LLMs, in the patent domain requires a deeper understanding of real-world patent examination. Our dataset is openly available at https://huggingface.co/datasets/LG-AI-Research/PANORAMA.

Research promotion enables researchers to share advanced knowledge with pertinent academic communities. The question-and-answer (QA) style articles are effective for researchers to promote their research by enabling readers to understand research on complex subjects. Recent advances in large language models (LLMs) have opened avenues for supporting researchers in creating QA-style articles for research promotion. However, without the authors’ involvement, these models may only partially capture the researcher’s intention and voice. We developed AQUA, a research probe that enables researchers to co-create QA-style articles with LLMs to promote their research papers. A user study (n=12) reveals that LLMs reduced authors’ burden and helped them understand the readers’ perspectives. Nevertheless, LLMs failed to capture the unique intent of the authors, and their automated generation discouraged authors from carefully revising their answers. Based on our findings, we discuss human-LLM interaction design to enable authors to create QA-style articles that reflect their intention.

The reliability of online dating algorithms has sparked considerable debate, particularly regarding skepticism about their excessive emphasis on evaluating and getting evaluated, which often overshadows the quest for authentic romantic connections. To understand the multifaceted influence of dating algorithms on end-users and explore avenues for algorithmic features considering the dynamics of human relationships, we conducted a mixed-methods study comprising in-depth interviews (N = 22) and a metaphoric co-design workshop (N = 12) with active users of online dating platforms. Interviews revealed that users perceive and respond to algorithmic evaluations with varied perceptions and behaviors, often expressing concerns about the emotional burden of constant self-presentation and the pursuit of quantitative assessments over genuine connections. In the design workshop, users envisioned desired algorithmic features to overcome investigated challenges, such as prioritizing personal values, tailored matchmaking, and support for personal growth in relationships. This research contributes to unraveling the complex dynamics of human-algorithm interaction in the context of online dating. By aligning algorithmic functions more closely with user desires and relationship goals, this study paves the way for more meaningful and authentic connections in the digital dating landscape.

Online dating is a space where autistic individuals can find romantic partners with reduced social demands. Autistic individuals are often expected to adapt their behaviors to the social norms underlying the online dating platform to appear as desirable romantic partners. However, given that their autistic traits can lead them to different expectations of dating, it is uncertain whether conforming their behaviors to the norm will guide them to the person they truly want. In this paper, we explored the perceptions and expectations of autistic adults in online dating through interviews and workshops. We found that autistic people desired to know whether they behaved according to the platform’s norms. Still, they expected to keep their unique characteristics rather than unconditionally conform to the norm. We conclude by providing suggestions for designing inclusive online dating experiences that could foster self-guided decisions of autistic users and embrace their unique characteristics.

The integration of generative AI in the field of education is actively being explored. In particular, ChatGPT has garnered significant interest, offering an opportunity to examine its effectiveness in English as a foreign language (EFL) education. To address this need, we present a novel learning platform called RECIPE (Revising an Essay with ChatGPT on an Interactive Platform for EFL learners). Our platform features two types of prompts that facilitate conversations between ChatGPT and students: (1) a hidden prompt for ChatGPT to take an EFL teacher role and (2) an open prompt for students to initiate a dialogue with a self-written summary of what they have learned. We deployed this platform for 213 undergraduate and graduate students enrolled in EFL writing courses and seven instructors. For this study, we collect students' interaction data from RECIPE, including students' perceptions and usage of the platform, and user scenarios are examined with the data. We also conduct a focus group interview with six students and an individual interview with one EFL instructor to explore design opportunities for leveraging generative AI models in the field of EFL education.

Current head-mounted displays enable users to explore virtual worlds by simply walking through them (i.e., real-walking VR). This led researchers to create haptic displays that can also simulate different types of elevation shapes. However, existing shape-changing floors are limited by their tabletop scale or the coarse resolution of the terrains they can display due to the limited number of actuators and low vertical resolution. To tackle this challenge, we introduce Elevate, a dynamic and walkable pin-array floor on which users can experience not only large variations in shapes but also the details of the underlying terrain. Our system achieves this by packing 1200 pins arranged on a 1.80 × 0.60m platform, in which each pin can be actuated to one of ten height levels (resolution: 15mm/level). To demonstrate its applicability, we present our haptic floor combined with four walkable applications and a user study that reported increased realism and enjoyment.