Recent Large Language Model (LLM) based AI can exhibit recognizable and measurable personality traits during conversations to improve user experience. However, as human understandings of their personality traits can be affected by their interaction partners’ traits, a potential risk is that AI traits may shape and bias users’ self-concept of their own traits. To explore the possibility, we conducted a randomized behavioral experiment. Our results indicate that after conversations about personal topics with an LLM-based AI chatbot using GPT-4o default personality traits, users’ self-concepts aligned with the AI’s measured personality traits. The longer the conversation, the greater the alignment. This alignment led to increased homogeneity in self-concepts among users. We also observed that the degree of self-concept alignment was positively associated with users’ conversation enjoyment. Our findings uncover how AI personality traits can shape users’ self-concepts through human-AI conversation, highlighting both risks and opportunities. We provide important design implications for developing more responsible and ethical AI systems.

Non-native speakers (NNSs) face significant language barriers in multilingual communication with native speakers (NSs). While AI-mediated communication (AIMC) tools offer efficient one-time assistance, they often overlook opportunities for NNSs’ continuous language acquisition. We introduce ChatLearn, an enhanced AIMC system that leverages NNSs’ communication difficulties as learning opportunities. Beyond comprehension and expression assistance, ChatLearn simultaneously captures NNSs’ language challenges, and subsequently provides them with spaced review as the conversation progresses. We conducted a mixed-methods study using a communication task with 43 NNS-NS pairs, after which ChatLearn NNSs recalled significantly more expressions than the baseline group, while there was no substantial decline in communication experience. Our findings highlight the value of contextual learning in NNS-NS communication, providing a new direction for AIMC systems that foster both immediate collaboration and continuous language development.

AI-assisted writing raises concerns about autonomy and ownership when benefiting writers. Personalization has been proposed as an effective solution while also risking writers’ reliance on AI and behavior shifting. For better personalization design, existing studies rely on interaction and information solely within the writing phase; however, few studies have examined how reading behaviors can inform personalized writing. This study investigates the effects of integrating reading highlights for personalization on AI-assisted writing. A between-subjects study with 46 participants revealed that the personalization condition encouraged participants to produce more highlights. However, highlighting unexpectedly shifted from a sense-making strategy to an instrumental act of "feeding the AI," leading to significant reliance on AI and declines in writers’ sense of autonomy, ownership, and self-credit. These findings indicate personalization risks in AI-assisted writing, emphasize the importance of personalization strategies, and provide design implications.

Individuals tend to apply preferences and beliefs as heuristics to effectively sift through the sheer amount of information available online. Such tendencies, however, often result in cognitive biases, which can skew judgment and open doors for manipulation. In this work, we investigate how individual and contextual factors lead to instances of confirmation bias when seeking, evaluating, and recalling polarising information. We conducted a lab study, in which we exposed participants to opinions on controversial issues through a Twitter-like news feed. We found that low-effortful thinking, strong political beliefs, and content conveying a strong issue amplify the occurrences of confirmation bias, leading to skewed information processing and recall. We discuss how the adverse effects of confirmation bias can be mitigated by taking bias-susceptibility into account. Specifically, social media platforms could aim to reduce strong expressions and integrate media literacy-building mechanisms, as low-effortful thinking styles and strong political beliefs render individuals especially susceptible to cognitive biases.

Computing systems are increasingly designed to adapt to users’ cognitive states and mental models. Yet, cognitive biases affect how humans form such models and, therefore, they can impact their interactions with computers. To better understand this interplay, we conducted a scoping review to chart how Human-Computer Interaction (HCI) researchers study cognitive biases. Our findings show that computing systems not only have the potential to induce and amplify cognitive biases but also can be designed to steer users’ behaviour and decision-making by capitalising on biases. We describe how HCI researchers develop algorithms and sensing methods to detect and quantify the effects of cognitive biases and discuss how we can use their understanding to inform system design. In this paper, we outline a research agenda for more theory-grounded research and highlight ethical issues when researching and designing computing systems with cognitive biases in mind as they affect real-world behaviour.

Cognitive biases have been shown to play a critical role in creating echo chambers and spreading misinformation. They undermine our ability to evaluate information and can influence our behaviour without our awareness. To allow the study of occurrences and effects of biases on information consumption behaviour, we explore indicators for cognitive biases in physiological and interaction data. Therefore, we conducted two experiments investigating how people experience statements that are congruent or divergent from their own ideological stance. We collected interaction data, eye tracking data, hemodynamic responses, and electrodermal activity while participants were exposed to ideologically tainted statements. Our results indicate that people spend more time processing statements that are incongruent with their own opinion. We detected differences in blood oxygenation levels between congruent and divergent opinions, a first step towards building systems to detect and quantify cognitive biases.

The performance of human workers can be fluctuated due to changes in the cognitive state during sustained work. Though past researches have made human performance monitoring possible by utilizing physiological signals, little attention has been paid to the context of office works. This research proposes a multimodal approach to estimate office task performance. A transcription typing experiment was conducted to simulate the real working environment while typing speed and error rate represented as performance metrics. Physiological data collected during the experiment, together with conventional machine learning algorithms showed feasibility to accurately predict two levels (good/bad) of task performance. More importantly, a comprehensive comparison between choices of modality suggests that using data from particular sources could gain predictive performance comparable to the complete set of modalities.