Millions of users form emotional attachments to AI companions like Character.AI, Replika, and ChatGPT. When these relationships end through model updates, safety interventions, or platform shutdowns, users receive no closure, reporting grief comparable to human loss. As regulations mandate protections for vulnerable users, discontinuation events will accelerate, yet no platform has implemented deliberate end-of-"life" design.

Marine ecosystems face unprecedented threats from climate change and plastic pollution, yet traditional environmental education often struggles to translate awareness into meaningful actions. This paper presents OceanChat, an interactive system leveraging large language models to create conversational AI agents represented as animated marine creatures, specifically a beluga whale, a jellyfish, and a seahorse, designed to promote pro-environmental behavior (PEB) and foster awareness through personalized dialogue. Through a between-subjects experiment (N=900), we compared three conditions: static scientific information, static character narratives, and interactive dialogue with AI-powered marine characters. Our analysis revealed that the Conversational Character Narrative condition significantly increased behavioral intentions and sustainable choice preferences compared to static approaches. The beluga whale character demonstrated consistently stronger emotional engagement across multiple measures, including perceived anthropomorphism and empathy. Our work extends research on sustainability interfaces facilitating PEB and offers design principles for creating emotionally resonant, intelligent AI characters.

Large Language Models (LLMs) are becoming increasingly ubiquitous in daily life, impacting decision-making across various domains. A substantial body of prior work has shown that individuals tend to evaluate positive predictions more favorably than negative ones—a phenomenon often referred to as the personal validation effect—across various non-AI prediction sources. Building on this foundation, we extend this well-established psychological effect to the context of LLM-based predictions, examining how prediction valence influences users’ perceptions when the source is an AI system. We investigate how positive AI-generated responses affect perceived validity, personalization, reliability, and usefulness of chatbot predictions, even when those predictions are fictitious and pre-scripted. In a study of 238 participants, positive predictions were perceived as significantly more valid (36% increase), personalized (42% increase), reliable (27% increase), and useful (22% increase) than negative predictions. These findings demonstrate that the personal validation effect persists in interactions with LLMs and underscore the substantial role of prediction valence in shaping user perceptions, with important implications for the design and deployment of AI systems across diverse applications.

Connecting with one’s future self has been shown to enhance decision-making, improve academic performance, promote positive health outcomes, and elevate subjective quality of life. Yet traditional interventions rely on imagination or static visualizations that may not be the most effective. AI-generated digital twins offer a new approach, enabling people to engage in dialogue with a personalized representation of themselves decades ahead. However, it remains unclear how presentation modality shapes their psychological impact. We report a randomized between-subjects study (n = 92) comparing three modalities of an AI-generated future self (text, voice, and a photorealistic talking avatar) against a generic AI control. Our system integrated age progression, voice cloning, and facial animation to create personalized digital twins. All personalized modalities significantly strengthened participants’ connection to their future selves, particularly in how vividly and positively they could imagine who they will become. Although the avatar produced the largest gain in vividness, effects were comparable across modalities. Instead, subjective interaction quality, especially perceived persuasiveness, realism, and engagement, strongly predicted gains in future self-continuity and affect, indicating that experiential quality matters more than interface form. Conversation analysis revealed modality-specific patterns, with text emphasizing instrumental career planning and voice-based interactions eliciting more existential reflection. These findings indicate that effective future-self interventions do not necessarily rely on resource-intensive architecture and can scale through less demanding formats. At the same time, they raise ethical considerations about the implications of persuasive AI that engages users’ own identities.

Millions of users now design personalized LLM-based chatbots through system prompts that shape their daily interactions, yet have limited ability to anticipate how these design choices will manifest as behaviors in deployment. This opacity is consequential: seemingly innocuous prompts can trigger excessive sycophancy, toxicity, or other undesirable traits, harming utility and raising safety concerns. To address this, we introduce an interface that enables neural transparency by exposing language model internals during the chatbot’s personality design. Our approach extracts behavioral trait vectors (empathy, toxicity, sycophancy, etc.) by calculating the differences in neural activations between contrastive system prompts that elicit opposing behaviors. We quantify a chatbot’s personality by projecting the system prompt’s final token activations onto these trait vectors to create persona scores, which are then normalized for cross-trait comparability and visualized using an interactive sunburst diagram. To evaluate this approach, we conducted an online user study (N = 80) to compare our neural transparency interface against a baseline chatbot interface without any form of transparency. Our analyses suggest that users systematically miscalibrated AI behavior: participants misjudged trait activations for 11 of 15 analyzable traits, motivating the need for transparency tools in everyday human-AI interaction. While our interface did not alter design iteration patterns, it significantly increased user trust and was enthusiastically received. Qualitative analysis revealed nuanced user experiences with the visualization, suggesting interface and interaction improvements for future work. This work offers a path for how mechanistic interpretability can be operationalized for non-technical users, establishing a method for safer, more aligned human-AI interactions.

CHI ’25, Yokohama, Japan

CHI ’25, Yokohama, Japan

Fostering students’ interests in learning is considered to have many positive downstream effects. Large language models have opened up new horizons for generating content tuned to one’s interests, yet it is unclear in what ways and to what extent this customization could have positive effects on learning. To explore this novel dimension, we conducted a between-subjects online study (n=272) featuring different variations of a generative AI vocabulary learning app that enables users to personalize their learning examples. Participants were randomly assigned to control (sentence sourced from pre-existing text) or experimental conditions (generated sentence or short story based on users’ text input). While we did not observe a difference in learning performance between the conditions, the analysis revealed that generative AI-driven context personalization positively affected learning motivation. We discuss how these results relate to previous findings and underscore their significance for the emerging field of using generative AI for personalized learning.

We introduce “Future You,” an interactive, brief, single-session, digital chat intervention designed to improve future self-continuity-the degree of connection an individual feels with a temporally distant future selfa characteristic that is positively related to mental health and wellbeing. Our system allows users to chat with a relatable yet AI-powered virtual version of their future selves that is tuned to their future goals and personal qualities. To make the conversation realistic, the system generates a “future memory”-a unique backstory for each user-that creates a throughline between the user's present age (between 18–30) and their life at age 60. The “Future You” character also adopts the persona of an age-progressed image of the user. In our preregistered study$(\mathrm{N}=344)$, we found that after a brief interaction with the “Future You” character, users reported significantly decreased anxiety and increased future self-continuity compared to control conditions. This is the first study successfully demonstrating the use of personalized AI-generated characters to improve users' future self-continuity and wellbeing.

This research full paper describes a web-based online learning platform that delivers financial literacy lessons via talking head videos of AI-generated personas with two additional core features: LLM-powered proactive chat-based question-and-answer interactivity, and personal choice of the AI instructor from a list of distinct personas. We conducted two comparative studies with a total of 233 Thai students aged 1825, which aim to 1) investigate the impact of interactivity and instructor selection on the learning experience, and 2) further explore the underlying factors at play with instructor selection by introducing AI instructors' backstories as an extra intervention. We found that enabling interactivity significantly enhanced learning motivation, perceived learning facilitation, engagement, and virtual instructors' humanness compared to the passive setting. Providing learners with a choice of AI instructors provided minimal additional benefit. However, the learner's feeling of relatedness toward the instructor is a significant positive predictor of learning motivation, positive emotion, and agent credibility, while goal alignment with the agent correlates with perceived learning facilitation, and admiration corresponds with perceived agent humanness. These findings underscore the potential of interactive virtual instructors-ones that interactively encourage learners to reflect on the teaching materials throughout the lesson through two-way interaction-in enhancing motivational and experiential aspects of remote education, even if they do not significantly impact comprehension, and the importance of promoting learner's relatedness and goal alignment with the agent in boosting other aspects of the learning experience.

Despite significant historical progress, discrimination and social stigma continue to impact the lives of LGBTQIA+ individuals. The use of AI-generated virtual characters offers a unique opportunity to facilitate advocacy by engaging individuals in simulated conversations that can foster understanding, education, and empathy. This paper explores the potential of AI simulations to help individuals practice LGBTQIA+ advocacy, while also acknowledging the need for ethical considerations and addressing concerns about oversimplification or perpetuation of stereotypes. By combining technological innovation with a commitment to inclusivity, we aim to contribute to the ongoing struggle for equality in both the legal framework and the hearts and minds of the community. We present a study evaluating virtual characters driven by generative conversational AI simulating the social interactions surrounding “coming out of the closet”, a rite of passage associated with LGBTQIA+ communities. In our study, virtual characters embodied as queer individuals engage with users in a text-based conversation simulation paired with visual representations. We investigate how the interactions between the virtual characters and a user influence the user’s comfort, confidence, empathy and sympathy. The AI simulation includes distinct visual personas deployed in a series of conditions. We present findings from our deployments involving 307 users. Finally, we discuss the design implications of our work on the potential future of embodied, self-actuated and openly LGBTQIA+ intelligent agents.

Critical thinking is an essential human skill. Despite the importance of critical thinking, research reveals that our reasoning ability suffers from personal biases and cognitive resource limitations, leading to potentially dangerous outcomes. This paper presents the novel idea of AI-framed Questioning that turns information relevant to the AI classification into questions to actively engage users’ thinking and scaffold their reasoning process. We conducted a study with 204 participants comparing the effects of AI-framed Questioning on a critical thinking task; discernment of logical validity of socially divisive statements. Our results show that compared to no feedback and even causal AI explanations of an always correct system, AI-framed Questioning significantly increase human discernment of logically flawed statements. Our experiment exemplifies a future style of Human-AI co-reasoning system, where the AI becomes a critical thinking stimulator rather than an information teller.

Every human culture has developed practices and rituals associated with remembering people of the past - be it for mourning, cultural preservation, or learning about historical events. In this paper, we present the concept of “Living Memories”: interactive digital mementos that are created from journals, letters and data that an individual have left behind. Like an interactive photograph, living memories can be talked to and asked questions, making accessing the knowledge, attitudes and past experiences of a person easily accessible. To demonstrate our concept, we created an AI-based system for generating living memories from any data source and implemented living memories of the three historical figures “Leonardo Da Vinci”, “Murasaki Shikibu”, and “Captain Robert Scott”. As a second key contribution, we present a novel metrics scheme for evaluating the accuracy of living memory architectures and show the accuracy of our pipeline to improve over baselines. Finally, we compare the user experience and learning effects of interacting with the living memory of Leonardo Da Vinci to reading his journal. Our results show that interacting with the living memory, in addition to simply reading a journal, increases learning effectiveness and motivation to learn about the character.

Video represents the majority of internet traffic today, driving a continual race between the generation of higher quality content, transmission of larger file sizes, and the development of network infrastructure. In addition, the recent COVID-19 pandemic fueled a surge in the use of video conferencing tools. Since videos take up considerable bandwidth ($\sim 100$Kbps to a few Mbps), improved video compression can have a substantial impact on network performance for live and pre-recorded content, providing broader access to multimedia content worldwide. We present a novel video compression pipeline, called Txt2Vid, which dramatically reduces data transmission rates by compressing webcam videos (“talking-head videos”) to a text transcript. The text is transmitted and decoded into a realistic reconstruction of the original video using recent advances in deep learning based voice cloning and lip syncing models. Our generative pipeline achieves two to three orders of magnitude reduction in the bitrate as compared to the standard audio-video codecs (encoders-decoders), while maintaining equivalent Quality-of-Experience based on a subjective evaluation by users ($n=242$) in an online study. The Txt2Vid framework opens up the potential for creating novel applications such as enabling audio-video communication during poor internet connectivity, or in remote terrains with limited bandwidth. The code for this work is available athttps://github.com/tpulkit/txt2vid.git.

This paper presents the results of a study with 134 participants to explore the effects of learning from an AI-generated virtual instructor that resembles a person one likes or admires. Given the important role instructors play in shaping learning experiences, as well as the recent surge in demand for online education, we investigate the potential for AI-generated instructors to motivate learning. Recent advances in generative AI have made it easy to create virtual instructors based on the likeness of a present-day, historical or fictional person, thereby enabling customization of video instructors based on the material, context and student. We found that while greater degrees of liking and admiration do not result in increased test scores, they can significantly improve students’ motivation towards learning, foster more positive emotions, and boost their appraisal of the AI-generated instructor as serving as an effective instructor.

Sense is a responsive garment that visualizes the global phenomenon of coral bleaching. Sense examines the current scientific shift in which technology converges with biology and embody the beauty, power and fragility of nature. This project explores aesthetic territory at the intersection of traditional textile techniques and wearable technologies. In contemporary fashion design, biomimicry inspiration provides alternative ways to express and communicate in the networked, hybrid physical-digital domain, and the urge of dying coral reefs.