An official website of the United States government
People form perceptions and interpretations of AI through external sources prior to their interaction with new technology. For example, shared anecdotes and media stories influence prior beliefs that may or may not accurately represent the true nature of AI systems. We hypothesize people's prior perceptions and beliefs will affect human-AI interactions and usage behaviors when using new applications. This paper presents a user experiment to explore the interplay between user's pre-existing beliefs about AI technology, individual differences, and previously established sources of cognitive bias from first impressions with an interactive AI application. We employed questionnaire measures as features to categorize users into profiles based on their prior beliefs and attitudes about technology. In addition, participants were assigned to one of two controlled conditions designed to evoke either positive or negative first impressions during an AI-assisted judgment task using an interactive application. The experiment and results provide empirical evidence that profiling users by surveying them on their prior beliefs and differences can be a beneficial approach for bias (and/or unanticipated usage) mitigation instead of seeking one-size-fits-all solutions.
more »
« less
Bias in AI Autocomplete Suggestions Leads to Attitude Shift on Societal Issues
AI technologies such as Large Language Models (LLMs) are increasingly used to make suggestions to autocomplete text as people write. Can these suggestions impact people’s writing and attitudes? In two large-scale preregistered experiments (N=2,582), we expose participants who are writing about important societal issues to biased AI-generated suggestions. The attitudes participants expressed in their writing and in a post-task survey converged towards the AI’s position. Yet, a majority of participants were unaware of the AI suggestions’ bias and their influence. Further, awareness of the task or of the AI’s bias, e.g. warning participants about potential bias before or after exposure to the treatment, did not mitigate the influence effect. Moreover, the AI’s influence is not fully explained by the additional information provided by the suggestions.
more »
« less
- Award ID(s):
- 1901151
- PAR ID:
- 10562887
- Publisher / Repository:
- PsyArXiv Preprints
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Given large language models’ (LLMs) increasing integration into workplace software, it is important to examine how biases in the models may impact workers. For example, stylistic biases in the language suggested by LLMs may cause feelings of alienation and result in increased labor for individuals or groups whose style does not match. We examine how such writer-style bias impacts inclusion, control, and ownership over the work when co-writing with LLMs. In an online experiment, participants wrote hypothetical job promotion requests using either hesitant or self-assured auto-complete suggestions from an LLM and reported their subsequent perceptions. We found that the style of the AI model did not impact perceived inclusion. However, individuals with higher perceived inclusion did perceive greater agency and ownership, an effect more strongly impacting participants of minoritized genders. Feelings of inclusion mitigated a loss of control and agency when accepting more AI suggestions.more » « less
-
Large language models (LLMs) are being increasingly integrated into everyday products and services, such as coding tools and writing assistants. As these embedded AI applications are deployed globally, there is a growing concern that the AI models underlying these applications prioritize Western values. This paper investigates what happens when a Western-centric AI model provides writing suggestions to users from a different cultural background. We conducted a cross-cultural controlled experiment with 118 participants from India and the United States who completed culturally grounded writing tasks with and without AI suggestions. Our analysis reveals that AI provided greater efficiency gains for Americans compared to Indians. Moreover, AI suggestions led Indian participants to adopt Western writing styles, altering not just what is written but also how it is written. These findings show that Western-centric AI models homogenize writing toward Western norms, diminishing nuances that differentiate cultural expression.more » « less
-
AI language technologies increasingly assist and expand human communication. While AI-mediated communication reduces human effort, its societal consequences are poorly understood. In this study, we investigate whether using an AI writing assistant in personal self-presentation changes how people talk about themselves. In an online experiment, we asked participants (N=200) to introduce themselves to others. An AI language assistant supported their writing by suggesting sentence completions. The language model generating suggestions was fine-tuned to preferably suggest either interest, work, or hospitality topics. We evaluate how the topic preference of a language model affected users’ topic choice by analyzing the topics participants discussed in their self-presentations. Our results suggest that AI language technologies may change the topics their users talk about. We discuss the need for a careful debate and evaluation of the topic priors built into AI language technologies.more » « less
-
User adoption of security and privacy (S&P) best practices remains low, despite sustained efforts by researchers and practitioners. Social influence is a proven method for guiding user S&P behavior, though most work has focused on studying peer influence, which is only possible with a known social graph. In a study of 104 Facebook users, we instead demonstrate that crowdsourced S&P suggestions are significantly influential. We also tested how reflective writing affected participants’ S&P decisions, with and without suggestions. With reflective writing, participants were less likely to accept suggestions — both social and Facebook default suggestions. Of particular note, when reflective writing participants were shown the Facebook default suggestion, they not only rejected it but also (unknowingly) configured their settings in accordance with expert recommendations. Our work suggests that both non-personal social influence and reflective writing can positively influence users’ S&P decisions, but have negative interactions.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Posted Content:
- https://doi.org/10.31234/osf.io/mhjn6
