With the rise of AI, more and more online content is generated by computers rather than people. Recent work has shown that AI-generated AirBnb profiles contributed to user mistrust only when users believed they saw a mixed set of human- and AI-generated profiles. However, AI results are based on statistical data, so not all results are considered equal. It remains unclear how indicators of algorithmic confidence like quantitative percentages or qualitative text descriptions (e.g., very confident) might affect user trust.
How do confidence indicators for AI-generated communication affect user trust?
Jakesch, M., French, M., Ma, X., Hancock, J., Mor Naaman. AI-Mediated Communication: How the Perception that Profile Text was Written by AI Affects Trustworthiness. CHI 2019. https://dl.acm.org/citation.cfm?id=3300469
Online trigger warnings caution users about content that some may find distressing. Yet, the presence of these trigger warnings currently depends entirely on the content creators’ decision to include them. This project seeks to design, build, and evaluate an automated trigger warning system (e.g., a browser plugin) that can selectively warn about or hide distressing content.
How can we automatically generate online trigger warnings for distressing content?
Kovacs, G., Wu, Z., Bernstein, M. Rotating Online Behavior Change Interventions Increases Effectiveness But Also Increases Attrition. CSCW 2018. https://hci.stanford.edu/publications/2018/habitlab/habitlab-cscw18.pdf
We interact with dozens of web pages daily and the number of people online continues to grow, making inclusive web design practices critical. Prior work investigating the sense of belonging in online spaces has shown that young women exposed to stereotypically masculine course webpages are negatively affected and experience more concern about others’ perception of their gender, as compared to those shown gender-neutral pages. The prior study suggests that gender biases can be triggered by web design, highlighting the need for inclusive user interface design for the web. However, the authors leave to future work to identify specific attributes of interest (including colorfulness, complexity, imagery, and language) that can be used to empirically determine the relative weight of such design factors for various biases in real computer science webpages. By isolating such factors as those listed above, this project aims to understand how specific attributes of design might impact and site viewer’s sense of belonging.
How do colorfulness, complexity, imagery, and/or language affect bias and belonging in web interfaces?
Metaxa, D., Wang, K., Landay, J.A., Hancock, J. Gender-Inclusive Design: Stereotypes and Biases in Web Interfaces. CHI 2018. https://dl.acm.org/citation.cfm?id=3173574.3174188
Whether in flipped classrooms or online learners, students increasingly rely on videos for learning. How do we know that these students are understanding the information being imparted? Prior work describes different forms of in-video prompting and their effect on attention. However, it does not explore the effect of these interventions on learning outcomes. This project will explore the impact of these different types of in-video prompts on users’ knowledge gain.
How do different kinds of in-video prompts affect learning?
Shin, H., Ko, E., Williams, J., Kim, J. Understanding the Effect of In-Video Prompting on Learners and Instructors. CHI 2018. http://juhokim.com/files/CHI2018-Prompting.pdf
Much of communication has shifted to digital forms: emails, texts, and direct messages. With that shift, comes new ways to moderate social interactions. Active now statuses let you know if someone is listening or when they were last available, typing indicators show when you are crafting a reply, and read receipts let your conversant know that you the message was received. In many ways these features mirror regular conversation, but they may not be received or interpreted in the same ways as their analog counterparts.
What is the effect of active now statuses, typing indicators, and read receipts on digital communication and meaning?
Hancock, J., Landrigan, C., Silver, C. Expressing Emotion in Text-based Communication. CHI 2007. https://sml.stanford.edu/ml/2007/05/hancock-chi-expressing-emotion-in-text.pdf