Pattie Maes

Caitlin Morris, Pattie Maes

AI chat tools are shifting problem-solving and brainstorming conversations away from colleagues and into private AI interactions, reducing the shared awareness that supports team coordination. We introduce InquiryBits, a system that shares minimal summaries of AI conversations within configurable trust boundaries, separating AI-only analysis from human-visible sharing. In a study with 80 professionals, we find that people are broadly willing to share these traces to support collaboration and avoid duplicating work - but only within bounded groups. Comfort drops sharply as audience expands beyond close teams; the level of detail shared matters less than who can see it, with a preference for more detail over less within trusted groups. These findings suggest that trust boundaries, more than information granularity, may be the most impactful design parameter.

Shruthi Shekar, Pat Pataranutaporn, Chethan Sarabu et al.

This paper presents a comprehensive analysis of how AI-generated medical responses are perceived and evaluated by non-experts. A total of 300 participants gave evaluations for medical responses that were either written by a medical doctor on an online healthcare platform, or generated by a large language model and labeled by physicians as having high or low accuracy. Results showed that participants could not effectively distinguish between AI-generated and Doctors' responses and demonstrated a preference for AI-generated responses, rating High Accuracy AI-generated responses as significantly more valid, trustworthy, and complete/satisfactory. Low Accuracy AI-generated responses on average performed very similar to Doctors' responses, if not more. Participants not only found these low-accuracy AI-generated responses to be valid, trustworthy, and complete/satisfactory but also indicated a high tendency to follow the potentially harmful medical advice and incorrectly seek unnecessary medical attention as a result of the response provided. This problematic reaction was comparable if not more to the reaction they displayed towards doctors' responses. This increased trust placed on inaccurate or inappropriate AI-generated medical advice can lead to misdiagnosis and harmful consequences for individuals seeking help. Further, participants were more trusting of High Accuracy AI-generated responses when told they were given by a doctor and experts rated AI-generated responses significantly higher when the source of the response was unknown. Both experts and non-experts exhibited bias, finding AI-generated responses to be more thorough and accurate than Doctors' responses but still valuing the involvement of a Doctor in the delivery of their medical advice. Ensuring AI systems are implemented with medical professionals should be the future of using AI for the delivery of medical advice.

Pat Pataranutaporn, Chayapatr Archiwaranguprok, Samantha W. T. Chan et al.

AI is increasingly used to enhance images and videos, both intentionally and unintentionally. As AI editing tools become more integrated into smartphones, users can modify or animate photos into realistic videos. This study examines the impact of AI-altered visuals on false memories--recollections of events that didn't occur or deviate from reality. In a pre-registered study, 200 participants were divided into four conditions of 50 each. Participants viewed original images, completed a filler task, then saw stimuli corresponding to their assigned condition: unedited images, AI-edited images, AI-generated videos, or AI-generated videos of AI-edited images. AI-edited visuals significantly increased false recollections, with AI-generated videos of AI-edited images having the strongest effect (2.05x compared to control). Confidence in false memories was also highest for this condition (1.19x compared to control). We discuss potential applications in HCI, such as therapeutic memory reframing, and challenges in ethical, legal, political, and societal domains.

Samantha Chan, Pat Pataranutaporn, Aditya Suri et al.

This study examines the impact of AI on human false memories -- recollections of events that did not occur or deviate from actual occurrences. It explores false memory induction through suggestive questioning in Human-AI interactions, simulating crime witness interviews. Four conditions were tested: control, survey-based, pre-scripted chatbot, and generative chatbot using a large language model (LLM). Participants (N=200) watched a crime video, then interacted with their assigned AI interviewer or survey, answering questions including five misleading ones. False memories were assessed immediately and after one week. Results show the generative chatbot condition significantly increased false memory formation, inducing over 3 times more immediate false memories than the control and 1.7 times more than the survey method. 36.4% of users' responses to the generative chatbot were misled through the interaction. After one week, the number of false memories induced by generative chatbots remained constant. However, confidence in these false memories remained higher than the control after one week. Moderating factors were explored: users who were less familiar with chatbots but more familiar with AI technology, and more interested in crime investigations, were more susceptible to false memories. These findings highlight the potential risks of using advanced AI in sensitive contexts, like police interviews, emphasizing the need for ethical considerations.

Valdemar Danry, Pat Pataranutaporn, Matthew Groh et al.

Advanced Artificial Intelligence (AI) systems, specifically large language models (LLMs), have the capability to generate not just misinformation, but also deceptive explanations that can justify and propagate false information and erode trust in the truth. We examined the impact of deceptive AI generated explanations on individuals' beliefs in a pre-registered online experiment with 23,840 observations from 1,192 participants. We found that in addition to being more persuasive than accurate and honest explanations, AI-generated deceptive explanations can significantly amplify belief in false news headlines and undermine true ones as compared to AI systems that simply classify the headline incorrectly as being true/false. Moreover, our results show that personal factors such as cognitive reflection and trust in AI do not necessarily protect individuals from these effects caused by deceptive AI generated explanations. Instead, our results show that the logical validity of AI generated deceptive explanations, that is whether the explanation has a causal effect on the truthfulness of the AI's classification, plays a critical role in countering their persuasiveness - with logically invalid explanations being deemed less credible. This underscores the importance of teaching logical reasoning and critical thinking skills to identify logically invalid arguments, fostering greater resilience against advanced AI-driven misinformation.

Cathy Mengying Fang, Sheer Karny, Chayapatr Archiwaranguprok et al.

Alignment research on large language models (LLMs) increasingly depends on understanding how these systems are used in everyday contexts. Yet naturalistic interaction data is difficult to access due to privacy constraints and platform control. We present AI-Wrapped, a prototype workflow for collecting naturalistic LLM chatbot usage data while providing participants with an immediate "wrapped"-style report on their usage statistics, top topics, and behavioral patterns. We report findings from an initial deployment with 82 U.S.-based adults across 48,495 conversations from their 2025 chat histories. Participants used LLMs for both instrumental and reflective purposes and had topics with emotional or existential themes. Some usage patterns reflect potential over-reliance or perfectionism. Heavy users showed comparatively more reflective exchanges than primarily transactional ones. Methodologically, even with zero data retention and PII removal, participants may remain hesitant to share chat data due to perceived privacy and judgment risks, underscoring the importance of transparent design when building measurement infrastructure for alignment research.

Eunhae Lee, Pat Pataranutaporn, Judith Amores et al.

Could belief in AI predictions be just another form of superstition? This study investigates psychological factors that influence belief in AI predictions about personal behavior, comparing it to belief in astrology- and personality-based predictions. Through an experiment with 238 participants, we examined how cognitive style, paranormal beliefs, AI attitudes, personality traits, and other factors affect perceived validity, reliability, usefulness, and personalization of predictions from different sources. Our findings reveal that belief in AI predictions is positively correlated with belief in predictions based on astrology and personality psychology. Notably, paranormal beliefs and positive attitudes about AI significantly increased perceived validity, reliability, usefulness, and personalization of AI predictions. Conscientiousness was negatively correlated with belief in predictions across all sources, and interest in the prediction topic increased believability across predictions. Surprisingly, we found no evidence that cognitive style has an impact on belief in fictitious AI-generated predictions. These results highlight the "rational superstition" phenomenon in AI, where belief is driven more by mental heuristics and intuition than critical evaluation. This research advances our understanding of the psychology of human-AI interaction, offering insights into designing and promoting AI systems that foster appropriate trust and skepticism, critical for responsible integration in an increasingly AI-driven world.

Minsol Michelle Kim, Daniel M. Low, David Lafond et al.

Breaking negative mental health cycles, including rumination and recurring regrets, requires reflection that translates awareness into behavioral change. Grounded in the Transtheoretical Model (TTM) and Gross's Emotion Regulation (ER) Process Model, we examine how Technologies Supporting Self-Reflection (TSR) bridge reflection and action. In a 15-day in-the-wild study (N = 20), participants used a voice-based journaling system to capture regrets and wishes and engaged in WhatIf-Planning, a novel structured reflection module integrating counterfactual thinking with if-then planning. Participants were randomized to either a free-form condition or a Gross-guided condition, which maps the five processes of Gross's ER model into explicit journaling prompts. We contribute: (1) a unified reflection-to-action TSR system that operationalizes the Preparation stage of TTM to bridge Contemplation and Action, and (2) triangulated empirical evidence from an in-the-wild journaling study that first operationalizes Gross's Process Model, revealing effects on coping flexibility and emotion regulation in daily life. Results show significant pre-post improvements in coping flexibility, indicating adaptive self-regulation across conditions, with the Gross-guided group generating more counterfactual alternatives, articulating concrete if-then action plans, and implementing more plans for self-driven change.

Anku Rani, Valdemar Danry, Paul Pu Liang et al.

Given the growing prevalence of fake information, including increasingly realistic AI-generated news, there is an urgent need to train people to better evaluate and detect misinformation. While interactions with AI have been shown to durably reduce people's beliefs in false information, it is unclear whether these interactions also teach people the skills to discern false information themselves. We conducted a month-long study where 67 participants classified news headline-image pairs as real or fake, discussed their assessments with an AI system, followed by an unassisted evaluation of unseen news items to measure accuracy before, during, and after AI assistance. While AI assistance produced immediate improvements during AI-assisted sessions (+21\% average), participants' unassisted performance on new items declined significantly by 15.3\% in week 4 compared to week 0. These results indicate that while AI may help immediately, it ultimately degrades long-term misinformation detection abilities

Chitralekha Gupta, Nadia Victoria Aritonang, Dixon Prem Daniel Rajendran et al.

Misinformation can spread rapidly in everyday conversation, where pausing to verify is not always possible. We envision a wearable system that bridges the timing gap between hearing a claim and forming a judgment. It uses ambient listening to detect verifiable claims, performs rapid web verification, and provides a subtle haptic nudge with a glanceable overview. A controlled study (N=34) simulated this approach and tested against a no-support baseline. Results show that instant, body-integrated feedback significantly improved real-time truth discernment and increased verification activity compared to unsupported fact-checking. However, it also introduced over-reliance when the system made errors, i.e. failed to flag false claims or flagged true claims as false. We contribute empirical evidence of improved discernment alongside insights into trust, effort, and user-system tensions in verification wearables.

Jonathan Albert Cohen, Kye Shimizu, Allen Song et al.

Robots in shared spaces often move in ways that are difficult for people to interpret, placing the burden on humans to adapt. High-DoF robots exhibit motion that people read as expressive, intentionally or not, making it important to understand how such cues are perceived. We present an online video study evaluating how different signaling modalities, expressive motion, lights, text, and audio, shape people's ability to understand a quadruped robot's upcoming navigation actions (Boston Dynamics Spot). Across four common scenarios, we measure how each modality influences humans' (1) accuracy in predicting the robot's next navigation action, (2) confidence in that prediction, and (3) trust in the robot to act safely. The study tests how expressive motions compare to explicit channels, whether aligned multimodal cues enhance interpretability, and how conflicting cues affect user confidence and trust. We contribute initial evidence on the relative effectiveness of implicit versus explicit signaling strategies.

Valdemar Danry, Javier Hernandez, Andrew Wilson et al.

Current LLM assistants are powerful at answering questions, but they have limited access to the behavioral context that reveals when and where a user is struggling. We present a gaze-grounded multimodal LLM assistant that uses egocentric video with gaze overlays to identify likely points of difficulty and target follow-up retrospective assistance. We instantiate this vision in a controlled study (n=36) comparing the gaze-aware AI assistant to a text-only LLM assistant. Compared to a conventional LLM assistant, the gaze-aware assistant was rated as significantly more accurate and personalized in its assessments of users' reading behavior and significantly improved people's ability to recall information. Users spoke significantly fewer words with the gaze-aware assistant, indicating more efficient interactions. Qualitative results underscored both perceived benefits in comprehension and challenges when interpretations of gaze behaviors were inaccurate. Our findings suggest that gaze-aware LLM assistants can reason about cognitive needs to improve cognitive outcomes of users.

Phoebe Chua, Cathy Mengying Fang, Takehiko Ohkawa et al.

Unlike spoken languages where the use of prosodic features to convey emotion is well studied, indicators of emotion in sign language remain poorly understood, creating communication barriers in critical settings. Sign languages present unique challenges as facial expressions and hand movements simultaneously serve both grammatical and emotional functions. To address this gap, we introduce EmoSign, the first sign video dataset containing sentiment and emotion labels for 200 American Sign Language (ASL) videos. We also collect open-ended descriptions of emotion cues. Annotations were done by 3 Deaf ASL signers with professional interpretation experience. Alongside the annotations, we include baseline models for sentiment and emotion classification. This dataset not only addresses a critical gap in existing sign language research but also establishes a new benchmark for understanding model capabilities in multimodal emotion recognition for sign languages. The dataset is made available at https://huggingface.co/datasets/catfang/emosign.

Manuel Cherep, Pranav M R, Pattie Maes et al.

The web is littered with images, once created for human consumption and now increasingly interpreted by agents using vision-language models (VLMs). These agents make visual decisions at scale, deciding what to click, recommend, or buy. Yet, we know little about the structure of their visual preferences. We introduce a framework for studying this by placing VLMs in controlled image-based choice tasks and systematically perturbing their inputs. Our key idea is to treat the agent's decision function as a latent visual utility that can be inferred through revealed preference: choices between systematically edited images. Starting from common images, such as product photos, we propose methods for visual prompt optimization, adapting text optimization methods to iteratively propose and apply visually plausible modifications using an image generation model (such as in composition, lighting, or background). We then evaluate which edits increase selection probability. Through large-scale experiments on frontier VLMs, we demonstrate that optimized edits significantly shift choice probabilities in head-to-head comparisons. We develop an automatic interpretability pipeline to explain these preferences, identifying consistent visual themes that drive selection. We argue that this approach offers a practical and efficient way to surface visual vulnerabilities, safety concerns that might otherwise be discovered implicitly in the wild, supporting more proactive auditing and governance of image-based AI agents.

Pulkit Tandon, Shubham Chandak, Pat Pataranutaporn et al.

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 (~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 at https://github.com/tpulkit/txt2vid.git.

Mina Khan, P Srivatsa, Advait Rane et al.

Data-efficiency and generalization are key challenges in deep learning and deep reinforcement learning as many models are trained on large-scale, domain-specific, and expensive-to-label datasets. Self-supervised models trained on large-scale uncurated datasets have shown successful transfer to diverse settings. We investigate using pretrained image representations and spatio-temporal attention for state representation learning in Atari. We also explore fine-tuning pretrained representations with self-supervised techniques, i.e., contrastive predictive coding, spatio-temporal contrastive learning, and augmentations. Our results show that pretrained representations are at par with state-of-the-art self-supervised methods trained on domain-specific data. Pretrained representations, thus, yield data and compute-efficient state representations. https://github.com/PAL-ML/PEARL_v1

Mina Khan, Pattie Maes

Egocentric visual context detection can support intelligence augmentation applications. We created a wearable system, called PAL, for wearable, personalized, and privacy-preserving egocentric visual context detection. PAL has a wearable device with a camera, heart-rate sensor, on-device deep learning, and audio input/output. PAL also has a mobile/web application for personalized context labeling. We used on-device deep learning models for generic object and face detection, low-shot custom face and context recognition (e.g., activities like brushing teeth), and custom context clustering (e.g., indoor locations). The models had over 80\% accuracy in in-the-wild contexts (~1000 images) and we tested PAL for intelligence augmentation applications like behavior change. We have made PAL is open-source to further support intelligence augmentation using personalized and privacy-preserving egocentric visual contexts.

Utkarsh Sarawgi, Rishab Khincha, Wazeer Zulfikar et al.

Reliability of machine learning (ML) systems is crucial in safety-critical applications such as healthcare, and uncertainty estimation is a widely researched method to highlight the confidence of ML systems in deployment. Sequential and parallel ensemble techniques have shown improved performance of ML systems in multi-modal settings by leveraging the feature sets together. We propose an uncertainty-aware boosting technique for multi-modal ensembling in order to focus on the data points with higher associated uncertainty estimates, rather than the ones with higher loss values. We evaluate this method on healthcare tasks related to Dementia and Parkinson's disease which involve real-world multi-modal speech and text data, wherein our method shows an improved performance. Additional analysis suggests that introducing uncertainty-awareness into the boosted ensembles decreases the overall entropy of the system, making it more robust to heteroscedasticity in the data, as well as better calibrating each of the modalities along with high quality prediction intervals. We open-source our entire codebase at https://github.com/usarawgi911/Uncertainty-aware-boosting

Rishab Khincha, Utkarsh Sarawgi, Wazeer Zulfikar et al.

The problem of missing data has been persistent for a long time and poses a major obstacle in machine learning and statistical data analysis. Past works in this field have tried using various data imputation techniques to fill in the missing data, or training neural networks (NNs) with the missing data. In this work, we propose a simple yet effective approach that clusters similar input features together using hierarchical clustering and then trains proportionately split neural networks with a joint loss. We evaluate this approach on a series of benchmark datasets and show promising improvements even with simple imputation techniques. We attribute this to learning through clusters of similar features in our model architecture. The source code is available at https://github.com/usarawgi911/Robustness-to-Missing-Features

Utkarsh Sarawgi, Wazeer Zulfikar, Rishab Khincha et al.

Reliability in Neural Networks (NNs) is crucial in safety-critical applications like healthcare, and uncertainty estimation is a widely researched method to highlight the confidence of NNs in deployment. In this work, we propose an uncertainty-aware boosting technique for multi-modal ensembling to predict Alzheimer's Dementia Severity. The propagation of uncertainty across acoustic, cognitive, and linguistic features produces an ensemble system robust to heteroscedasticity in the data. Weighing the different modalities based on the uncertainty estimates, we experiment on the benchmark ADReSS dataset, a subject-independent and balanced dataset, to show that our method outperforms the state-of-the-art methods while also reducing the overall entropy of the system. This work aims to encourage fair and aware models. The source code is available at https://github.com/wazeerzulfikar/alzheimers-dementia

Utkarsh Sarawgi, Wazeer Zulfikar, Rishab Khincha et al.

Understanding and quantifying uncertainty in black box Neural Networks (NNs) is critical when deployed in real-world settings such as healthcare. Recent works using Bayesian and non-Bayesian methods have shown how a unified predictive uncertainty can be modelled for NNs. Decomposing this uncertainty to disentangle the granular sources of heteroscedasticity in data provides rich information about its underlying causes. We propose a conceptually simple non-Bayesian approach, deep split ensemble, to disentangle the predictive uncertainties using a multivariate Gaussian mixture model. The NNs are trained with clusters of input features, for uncertainty estimates per cluster. We evaluate our approach on a series of benchmark regression datasets, while also comparing with unified uncertainty methods. Extensive analyses using dataset shits and empirical rule highlight our inherently well-calibrated models. Our work further demonstrates its applicability in a multi-modal setting using a benchmark Alzheimer's dataset and also shows how deep split ensembles can highlight hidden modality-specific biases. The minimal changes required to NNs and the training procedure, and the high flexibility to group features into clusters makes it readily deployable and useful. The source code is available at https://github.com/wazeerzulfikar/deep-split-ensembles

Utkarsh Sarawgi, Wazeer Zulfikar, Nouran Soliman et al.

Alzheimer's disease is estimated to affect around 50 million people worldwide and is rising rapidly, with a global economic burden of nearly a trillion dollars. This calls for scalable, cost-effective, and robust methods for detection of Alzheimer's dementia (AD). We present a novel architecture that leverages acoustic, cognitive, and linguistic features to form a multimodal ensemble system. It uses specialized artificial neural networks with temporal characteristics to detect AD and its severity, which is reflected through Mini-Mental State Exam (MMSE) scores. We first evaluate it on the ADReSS challenge dataset, which is a subject-independent and balanced dataset matched for age and gender to mitigate biases, and is available through DementiaBank. Our system achieves state-of-the-art test accuracy, precision, recall, and F1-score of 83.3% each for AD classification, and state-of-the-art test root mean squared error (RMSE) of 4.60 for MMSE score regression. To the best of our knowledge, the system further achieves state-of-the-art AD classification accuracy of 88.0% when evaluated on the full benchmark DementiaBank Pitt database. Our work highlights the applicability and transferability of spontaneous speech to produce a robust inductive transfer learning model, and demonstrates generalizability through a task-agnostic feature-space. The source code is available at https://github.com/wazeerzulfikar/alzheimers-dementia

Mina Khan, Glenn Fernandes, Utkarsh Sarawgi et al.

Personalized Active Learner (PAL) is a wearable system for real-time, personalized, and context-aware health and cognition support. PAL's system consists of a wearable device, mobile app, cloud database, data visualization web app, and machine learning server. PAL's wearable device uses multi-modal sensors (camera, microphone, heart-rate) with on-device machine learning and open-ear audio output to provide real-time and context-aware cognitive, behavioral and psychological interventions. PAL also allows users to track the long-term correlations between their activities and physiological states to make well-informed lifestyle decisions. In this paper, we present and open-source PAL's system so that people can use it for health and cognition support applications. We also open-source three fully-developed example applications using PAL for face-based memory augmentation, contextual language learning, and heart-rate-based psychological support. PAL's flexible, modular and extensible platform combines trends in data-driven medicine, mobile psychology, and cognitive enhancement to support data-driven and empowering health and cognition applications.

Nataliya Kosmyna, Eugene Hauptmann, Ye Tong Yuan et al.

This study explores the neural and behavioral consequences of LLM-assisted essay writing. Participants were divided into three groups: LLM, Search Engine, and Brain-only (no tools). Each completed three sessions under the same condition. In a fourth session, LLM users were reassigned to Brain-only group (LLM-to-Brain), and Brain-only users were reassigned to LLM condition (Brain-to-LLM). A total of 54 participants took part in Sessions 1-3, with 18 completing session 4. We used electroencephalography (EEG) to assess cognitive load during essay writing, and analyzed essays using NLP, as well as scoring essays with the help from human teachers and an AI judge. Across groups, NERs, n-gram patterns, and topic ontology showed within-group homogeneity. EEG revealed significant differences in brain connectivity: Brain-only participants exhibited the strongest, most distributed networks; Search Engine users showed moderate engagement; and LLM users displayed the weakest connectivity. Cognitive activity scaled down in relation to external tool use. In session 4, LLM-to-Brain participants showed reduced alpha and beta connectivity, indicating under-engagement. Brain-to-LLM users exhibited higher memory recall and activation of occipito-parietal and prefrontal areas, similar to Search Engine users. Self-reported ownership of essays was the lowest in the LLM group and the highest in the Brain-only group. LLM users also struggled to accurately quote their own work. While LLMs offer immediate convenience, our findings highlight potential cognitive costs. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels. These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI's role in learning.

Pat Pataranutaporn, Kavin Winson, Peggy Yin et al.

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 self--a 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 "synthetic 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's present self. After a brief interaction with the "Future You" character, users reported decreased anxiety, and increased future self-continuity. This is the first study successfully demonstrating the use of personalized AI-generated characters to improve users' future self-continuity and wellbeing.

Jason Phang, Michael Lampe, Lama Ahmad et al.

As AI chatbots see increased adoption and integration into everyday life, questions have been raised about the potential impact of human-like or anthropomorphic AI on users. In this work, we investigate the extent to which interactions with ChatGPT (with a focus on Advanced Voice Mode) may impact users' emotional well-being, behaviors and experiences through two parallel studies. To study the affective use of AI chatbots, we perform large-scale automated analysis of ChatGPT platform usage in a privacy-preserving manner, analyzing over 3 million conversations for affective cues and surveying over 4,000 users on their perceptions of ChatGPT. To investigate whether there is a relationship between model usage and emotional well-being, we conduct an Institutional Review Board (IRB)-approved randomized controlled trial (RCT) on close to 1,000 participants over 28 days, examining changes in their emotional well-being as they interact with ChatGPT under different experimental settings. In both on-platform data analysis and the RCT, we observe that very high usage correlates with increased self-reported indicators of dependence. From our RCT, we find that the impact of voice-based interactions on emotional well-being to be highly nuanced, and influenced by factors such as the user's initial emotional state and total usage duration. Overall, our analysis reveals that a small number of users are responsible for a disproportionate share of the most affective cues.

Dünya Baradari, Nataliya Kosmyna, Oscar Petrov et al.

Generative AI is transforming education by enabling personalized, on-demand learning experiences. However, AI tutors lack the ability to assess a learner's cognitive state in real time, limiting their adaptability. Meanwhile, electroencephalography (EEG)-based neuroadaptive systems have successfully enhanced engagement by dynamically adjusting learning content. This paper presents NeuroChat, a proof-of-concept neuroadaptive AI tutor that integrates real-time EEG-based engagement tracking with generative AI. NeuroChat continuously monitors a learner's cognitive engagement and dynamically adjusts content complexity, response style, and pacing using a closed-loop system. We evaluate this approach in a pilot study (n=24), comparing NeuroChat to a standard LLM-based chatbot. Results indicate that NeuroChat enhances cognitive and subjective engagement but does not show an immediate effect on learning outcomes. These findings demonstrate the feasibility of real-time cognitive feedback in LLMs, highlighting new directions for adaptive learning, AI tutoring, and human-AI interaction.

Philipp Hugenroth, Valdemar Danry, Pattie Maes

As Large Language Models (LLMs) increasingly automate writing tasks, there is a growing risk of cognitive deskilling where users offload critical thinking to the system. To address this, we introduce Critical Inker, a writing tool designed to scaffold critical reflection during writing through logical analysis and socratic feedback. We present two methods: (1) A Socratic chatbot using questions to help them realize and fix logical errors in their writing and (2) Visual Feedback, which highlights logical errors in the text without dialog. We detail the technical implementation of the system and evaluate its argument extraction and logical validity accuracy. Our evaluation shows a 91.2% argument overlap with ground truth argument annotations and 87% validity accuracy. Finally, we conducted a small-scale pilot and discuss early qualitative results.

Caitlin Morris, Pattie Maes

As AI systems increasingly take on instructional roles - providing feedback, guiding practice, evaluating work - a fundamental question emerges: does it matter to learners who they believe is on the other side? We investigated this using a three-condition experiment (N=148) in which participants completed a creative coding tutorial and received feedback generated by the same large language model, attributed to either an AI system (with instant or delayed delivery) or a human teaching assistant (with matched delayed delivery). This three-condition design separates the effect of source attribution from the confound of delivery timing, which prior studies have not controlled. Source attribution and timing had distinct effects on different outcomes: participants who believed the human attribution spent more time on task than those receiving equivalently timed AI-attributed feedback (d=0.61, p=.013, uncorrected), while the delivery delay independently increased output complexity without affecting time measures. An exploratory analysis revealed that 46% of participants in the human-attributed condition did not believe the attribution, and these participants showed worse outcomes than those receiving transparent AI feedback (code complexity d=0.77, p=.003; time on task d=0.70, p=.007). These findings suggest that believed human presence may carry motivational value, but that this value depends on credibility. For computing educators, transparent AI attribution may be the lower-risk default in contexts where human attribution would not be credible.

Nikhil Singh, Manuel Cherep, Pattie Maes · mit

The fidelity with which neural networks can now generate content such as music presents a scientific opportunity: these systems appear to have learned implicit theories of such content's structure through statistical learning alone. This offers a potentially new lens on theories of human-generated media. When internal representations align with traditional constructs (e.g. chord progressions in music), they show how such categories can emerge from statistical regularities; when they diverge, they expose limits of existing frameworks and patterns we may have overlooked but that nonetheless carry explanatory power. In this paper, focusing on music generators, we introduce a method for discovering interpretable concepts using sparse autoencoders (SAEs), extracting interpretable features from the residual stream of a transformer model. We make this approach scalable and evaluable using automated labeling and validation pipelines. Our results reveal both familiar musical concepts and coherent but uncodified patterns lacking clear counterparts in theory or language. As an extension, we show such concepts can be used to steer model generations. Beyond improving model transparency, our work provides an empirical tool for uncovering organizing principles that have eluded traditional methods of analysis and synthesis.

Wazeer Zulfikar, Treyden Chiaravalloti, Jocelyn Shen et al.

People inherently use experiences of their past while imagining their future, a capability that plays a crucial role in mental health. Resonance is an AI-powered journaling tool designed to augment this ability by offering AI-generated, action-oriented suggestions for future activities based on the user's own past memories. Suggestions are offered when a new memory is logged and are followed by a prompt for the user to imagine carrying out the suggestion. In a two-week randomized controlled study (N=55), we found that using Resonance significantly improved mental health outcomes, reducing the users' PHQ8 scores, a measure of current depression, and increasing their daily positive affect, particularly when they would likely act on the suggestion. Notably, the effectiveness of the suggestions was higher when they were personal, novel, and referenced the user's logged memories. Finally, through open-ended feedback, we discuss the factors that encouraged or hindered the use of the tool.

Alexander Doudkin, Pat Pataranutaporn, Pattie Maes

Pro-environmental behavior (PEB) is vital to combat climate change, yet turning awareness into intention and action remains elusive. We explore large language models (LLMs) as tools to promote PEB, comparing their impact across 3,200 participants: real humans (n=1,200), simulated humans based on actual participant data (n=1,200), and fully synthetic personas (n=1,200). All three participant groups faced personalized or standard chatbots, or static statements, employing four persuasion strategies (moral foundations, future self-continuity, action orientation, or "freestyle" chosen by the LLM). Results reveal a "synthetic persuasion paradox": synthetic and simulated agents significantly affect their post-intervention PEB stance, while human responses barely shift. Simulated participants better approximate human trends but still overestimate effects. This disconnect underscores LLM's potential for pre-evaluating PEB interventions but warns of its limits in predicting real-world behavior. We call for refined synthetic modeling and sustained and extended human trials to align conversational AI's promise with tangible sustainability outcomes.

Pat Pataranutaporn, Alexander Doudkin, Pattie Maes

Marine ecosystems face unprecedented threats from climate change and plastic pollution, yet traditional environmental education often struggles to translate awareness into sustained behavioral change. 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 environmental behavior (PEB) and foster awareness through personalized dialogue. Through a between-subjects experiment (N=900), we compared three conditions: (1) Static Scientific Information, providing conventional environmental education through text and images; (2) Static Character Narrative, featuring first-person storytelling from 3D-rendered marine creatures; and (3) Conversational Character Narrative, enabling real-time 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. However, impacts on deeper measures like climate policy support and psychological distance were limited, highlighting the complexity of shifting entrenched beliefs. Our work extends research on sustainability interfaces facilitating PEB and offers design principles for creating emotionally resonant, context-aware AI characters. By balancing anthropomorphism with species authenticity, OceanChat demonstrates how interactive narratives can bridge the gap between environmental knowledge and real-world behavior change.

Constanze Albrecht, Chayapatr Archiwaranguprok, Rachel Poonsiriwong et al.

What if users could meet their future selves today? AI-generated future selves simulate meaningful encounters with a digital twin decades in the future. As AI systems advance, combining cloned voices, age-progressed facial rendering, and autobiographical narratives, a central question emerges: Does the modality of these future selves alter their psychological and affective impact? How might a text-based chatbot, a voice-only system, or a photorealistic avatar shape present-day decisions and our feeling of connection to the future? We report a randomized controlled study (N=92) evaluating three modalities of AI-generated future selves (text, voice, avatar) against a neutral control condition. We also report a systematic model evaluation between Claude 4 and three other Large Language Models (LLMs), assessing Claude 4 across psychological and interaction dimensions and establishing conversational AI quality as a critical determinant of intervention effectiveness. All personalized modalities strengthened Future Self-Continuity (FSC), emotional well-being, and motivation compared to control, with avatar producing the largest vividness gains, yet with no significant differences between formats. Interaction quality metrics, particularly persuasiveness, realism, and user engagement, emerged as robust predictors of psychological and affective outcomes, indicating that how compelling the interaction feels matters more than the form it takes. Content analysis found thematic patterns: text emphasized career planning, while voice and avatar facilitated personal reflection. Claude 4 outperformed ChatGPT 3.5, Llama 4, and Qwen 3 in enhancing psychological, affective, and FSC outcomes.

Manuel Cherep, Chengtian Ma, Abigail Xu et al.

Environments built for people are increasingly operated by a new class of economic actors: LLM-powered software agents making decisions on our behalf. These decisions range from our purchases to travel plans to medical treatment selection. Current evaluations of these agents largely focus on task competence, but we argue for a deeper assessment: how these agents choose when faced with realistic decisions. We introduce ABxLab, a framework for systematically probing agentic choice through controlled manipulations of option attributes and persuasive cues. We apply this to a realistic web-based shopping environment, where we vary prices, ratings, and psychological nudges, all of which are factors long known to shape human choice. We find that agent decisions shift predictably and substantially in response, revealing that agents are strongly biased choosers even without being subject to the cognitive constraints that shape human biases. This susceptibility reveals both risk and opportunity: risk, because agentic consumers may inherit and amplify human biases; opportunity, because consumer choice provides a powerful testbed for a behavioral science of AI agents, just as it has for the study of human behavior. We release our framework as an open benchmark for rigorous, scalable evaluation of agent decision-making.

Haojun Zhuang, Dünya Baradari, Nataliya Kosmyna et al.

Humans regularly navigate an overwhelming amount of information via text media, whether reading articles, browsing social media, or interacting with chatbots. Confusion naturally arises when new information conflicts with or exceeds a reader's comprehension or prior knowledge, posing a challenge for learning. In this study, we present a multimodal investigation of reading-induced confusion using EEG and eye tracking. We collected neural and gaze data from 11 adult participants as they read short paragraphs sampled from diverse, real-world sources. By isolating the N400 event-related potential (ERP), a well-established neural marker of semantic incongruence, and integrating behavioral markers from eye tracking, we provide a detailed analysis of the neural and behavioral correlates of confusion during naturalistic reading. Using machine learning, we show that multimodal (EEG + eye tracking) models improve classification accuracy by 4-22% over unimodal baselines, reaching an average weighted participant accuracy of 77.3% and a best accuracy of 89.6%. Our results highlight the dominance of the brain's temporal regions in these neural signatures of confusion, suggesting avenues for wearable, low-electrode brain-computer interfaces (BCI) for real-time monitoring. These findings lay the foundation for developing adaptive systems that dynamically detect and respond to user confusion, with potential applications in personalized learning, human-computer interaction, and accessibility.

Pat Pataranutaporn, Nattavudh Powdthavee, Chayapatr Achiwaranguprok et al.

This study examines the potential of large language models (LLMs) to augment the academic peer review process by reliably evaluating the quality of economics research without introducing systematic bias. We conduct one of the first large-scale experimental assessments of four LLMs (GPT-4o, Claude 3.5, Gemma 3, and LLaMA 3.3) across two complementary experiments. In the first, we use nonparametric binscatter and linear regression techniques to analyze over 29,000 evaluations of 1,220 anonymized papers drawn from 110 economics journals excluded from the training data of current LLMs, along with a set of AI-generated submissions. The results show that LLMs consistently distinguish between higher- and lower-quality research based solely on textual content, producing quality gradients that closely align with established journal prestige measures. Claude and Gemma perform exceptionally well in capturing these gradients, while GPT excels in detecting AI-generated content. The second experiment comprises 8,910 evaluations designed to assess whether LLMs replicate human like biases in single blind reviews. By systematically varying author gender, institutional affiliation, and academic prominence across 330 papers, we find that GPT, Gemma, and LLaMA assign significantly higher ratings to submissions from top male authors and elite institutions relative to the same papers presented anonymously. These results emphasize the importance of excluding author-identifying information when deploying LLMs in editorial screening. Overall, our findings provide compelling evidence and practical guidance for integrating LLMs into peer review to enhance efficiency, improve accuracy, and promote equity in the publication process of economics research.

Pat Pataranutaporn, Nattavudh Powdthavee, Pattie Maes

Surnames often convey implicit markers of social status, wealth, and lineage, shaping perceptions in ways that can perpetuate systemic biases and intergenerational inequality. This study is the first of its kind to investigate whether and how surnames influence AI-driven decision-making, focusing on their effects across key areas such as hiring recommendations, leadership appointments, and loan approvals. Using 72,000 evaluations of 600 surnames from the United States and Thailand, two countries with distinct sociohistorical contexts and surname conventions, we classify names into four categories: Rich, Legacy, Normal, and phonetically similar Variant groups. Our findings show that elite surnames consistently increase AI-generated perceptions of power, intelligence, and wealth, which in turn influence AI-driven decisions in high-stakes contexts. Mediation analysis reveals perceived intelligence as a key mechanism through which surname biases influence AI decision-making process. While providing objective qualifications alongside surnames mitigates most of these biases, it does not eliminate them entirely, especially in contexts where candidate credentials are low. These findings highlight the need for fairness-aware algorithms and robust policy measures to prevent AI systems from reinforcing systemic inequalities tied to surnames, an often-overlooked bias compared to more salient characteristics such as race and gender. Our work calls for a critical reassessment of algorithmic accountability and its broader societal impact, particularly in systems designed to uphold meritocratic principles while counteracting the perpetuation of intergenerational privilege.

Mina Khan, Zeel Patel, Kathryn Wantlin et al.

Technology based screentime, the time an individual spends engaging with their computer or cell phone, has increased exponentially over the past decade, but perhaps most alarmingly amidst the COVID-19 pandemic. Although many software based interventions exist to reduce screentime, users report a variety of issues relating to the timing of the intervention, the strictness of the tool, and its ability to encourage organic, long-term habit formation. We develop guidelines for the design of behaviour intervention software by conducting a survey to investigate three research questions and further inform the mechanisms of computer-related behaviour change applications. RQ1: What do people want to change and why/how? RQ2: What applications do people use or have used, why do they work or not, and what additional support is desired? RQ3: What are helpful/unhelpful computer breaks and why? Our survey had 68 participants and three key findings. First, time management is a primary concern, but emotional and physical side-effects are equally important. Second, site blockers, self-trackers, and timers are commonly used, but they are ineffective as they are easy-to-ignore and not personalized. Third, away-from-computer breaks, especially involving physical activity, are helpful, whereas on-screen breaks are unhelpful, especially when they are long, because they are not refreshing. We recommend personalized and closed-loop computer-usage behaviour change support and especially encouraging off-the-computer screentime breaks.

Mina Khan, P Srivatsa, Advait Rane et al.

Self-supervised or weakly supervised models trained on large-scale datasets have shown sample-efficient transfer to diverse datasets in few-shot settings. We consider how upstream pretrained models can be leveraged for downstream few-shot, multilabel, and continual learning tasks. Our model CLIPPER (CLIP PERsonalized) uses image representations from CLIP, a large-scale image representation learning model trained using weak natural language supervision. We developed a technique, called Multi-label Weight Imprinting (MWI), for multi-label, continual, and few-shot learning, and CLIPPER uses MWI with image representations from CLIP. We evaluated CLIPPER on 10 single-label and 5 multi-label datasets. Our model shows robust and competitive performance, and we set new benchmarks for few-shot, multi-label, and continual learning. Our lightweight technique is also compute-efficient and enables privacy-preserving applications as the data is not sent to the upstream model for fine-tuning.

Abhay Koushik, Judith Amores, Pattie Maes

We present the first real-time sleep staging system that uses deep learning without the need for servers in a smartphone application for a wearable EEG. We employ real-time adaptation of a single channel Electroencephalography (EEG) to infer from a Time-Distributed 1-D Deep Convolutional Neural Network. Polysomnography (PSG)-the gold standard for sleep staging, requires a human scorer and is both complex and resource-intensive. Our work demonstrates an end-to-end on-smartphone pipeline that can infer sleep stages in just single 30-second epochs, with an overall accuracy of 83.5% on 20-fold cross validation for five-class classification of sleep stages using the open Sleep-EDF dataset.