LXLarry Zhiming Xu

Larry Zhiming Xu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Larry Zhiming Xu has a total of 11 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    More Than Use and Cheating: A Developmental Study of High School Students’ Meaning- Making About AI

    by Gabriel Velez, Aubrie Robinson, Larry Zhiming Xu, Sahil Bagai

    Artificial intelligence (AI) has become increasingly embedded in adolescents’ educational experiences, yet little research has examined how young people make sense of AI and its implications for their development. Guided by the Phenomenological Variant of Ecological Systems Theory (PVEST), this qualitative study explored how high school students understand AI in education and how those understandings relate to their identities as students. Semi-structured interviews were conducted with 18 students from 11 U.S. high schools and analyzed using reflexive thematic analysis. Students commonly framed AI as a neutral tool whose value depended on how it was used. They described their own AI use as supporting learning and coping with academic stress while characterizing peers’ use as more likely to involve cheating. Participants also expressed fears about dependency, diminished learning, and future employment. Finally, they perceived AI as altering learning processes and contributing to mistrust and distance in student– teacher relationships. Findings highlight AI as a developmental and relational phenomenon rooted in identity processes central to adolescence.

  • Open Access

    LLMs Can Amplify Emotion While Preserving Analytic Style in AI-mediated Communication

    by Samuel Mazzone, Jonah Harlan, K M Sajjadul Islam, Larry Zhiming Xu, Terence Ow

    The ubiquity of generative AI (GenAI) requires social computing scholarship to examine how such extensive AI mediation may (re)shape the emotional and analytic qualities of human communication, both of which are central to audience engagement, trust, and information integrity. This study conducts a large-scale evaluation of 11 custom and fine-tuned LLMs, including DeepSeek, Llama (Meta), Mistral, and Gemma (Google), by asking them to rewrite the complete set of content, published by a local news outlet over a 12-year span (2011–2023). Using transformer-based sentiment models (i.e., RoBERTa fine-tuned on GoEmotions) and lexicon-based measures (i.e., LIWC, NRCLex), we compare the emotional expressiveness and analytic style of human-written versus AI-adapted content across short-form (titles/headlines) and long-form communication (article bodies/content). Analyzing 3,623 original news articles, nearly 40,000 primary-prompt AI-generated rewrites, and an additional generic-prompt ablation set, we estimate linear mixed-effects models that account for article-level clustering and control for model and prompt heterogeneity, communication form, verbosity, and readability. Results show that, under the corpus and prompting conditions examined here, LLMs tend to intensify emotional expression relative to human-written texts. Meanwhile, LLM rewrites tend to preserve, and in many cases increase, linguistic markers associated with analytic style relative to human benchmarks. Interestingly, even under the simpler generic prompt, LLMs in our sample display a shift toward more positive emotional categories, including joy, excitement, admiration, optimism, and gratitude, whereas human writers in this corpus display a broader affective range and more nuanced expression. To assess risks of clickbait-like headline framing, we fine-tune a BERT model for clickbait detection and examine linguistic cues in titles and headlines, yet we find minimal evidence that LLMs’ emotional amplification devolves into sensational and hyperbolic framing. Lastly, to enhance validity, we conduct a blinded human evaluation study on a randomly sampled subset of the corpus. Across 17 small experiments, human judges ( n = 170) independently evaluated analytic style, reasoning depth, and clickbait characteristics. These human evaluations provide convergent evidence that AI-generated texts were perceived as more analytically structured and slightly higher in perceived reasoning depth, while AI-generated headlines were also rated as less clickbait-oriented. Overall, our findings suggest that GenAI and LLMs have the potential to enhance the affective intensity of communication without necessarily compromising its informational quality. However, as human communication becomes increasingly and inevitably AI-mediated, current LLMs may shift the affective register, narrow the emotional spectrum, and introduce tonal biases in public communication that could subtly influence how information is disseminated, and how language is traditionally interpreted.

  • Open Access

    When Confidence Shifts Toward the Tool: AI Appropriation, Reliance Calibration, and School Guidance in High School Academic Work

    by Larry Zhiming Xu, Gabriel Velez, Aubrie Robinson

    As AI permeates education, adoption, by itself, no longer tells us much about the evolving role the technology plays. To provide new insights, we surveyed 1,444 students in a network of college-preparatory high schools in the United States. We draw from technology management literature and focus on appropriation (i.e., the purposes students assign to AI) and reliance calibration (i.e., the gap between students’ confidence in AI and that in themselves). Regarding appropriation, we found learning-oriented and shortcut-oriented purposes often co-occurred, suggesting that students who asked AI to explain a concept or outline a paper were often the same students who asked it for homework answers or full drafts. This complexity in appropriation challenges the simple dichotomy between “good” and “bad” AI users. In terms of reliance, use frequency was negatively associated with self-confidence without AI and positively associated with reliance calibration, indicating a potentially widening reliance gap. In hierarchical regressions with school-clustered standard errors, use frequency, breadth of purposes, and optimism about AI predicted this gap, while perceived school guidance, surprisingly, did not. Notably, students most often reported receiving general guidelines from schools, but their largest unmet demand was teacher-guided practice. Based on the results, we argue that AI reliance calibration captures the new dynamic relationship between technological capabilities and students’ self-perceptions. Therefore, it should draw the attention from education researchers and schools to understand how AI is used in today’s educational context.

  • When AI Becomes Its Own Biggest Fan: Self-Preference Bias in AI-Assisted Peer Review

    by Jianzheng Shi, Jungmin Lee, Larry Zhiming Xu, Terence T. Ow, Yue Wang

    This study examined whether large language models exhibit self-preference bias, a tendency to favor texts generated by themselves over those produced by other models, in AI-assisted peer review. Using a controlled review simulation, GPT-4, GPT-3, and LLaMA were used to generate and evaluate 60 academic manuscripts under a common rubric covering clarity, relevance, originality, and methodological appropriateness. This design enabled direct comparison between self-reviews (a model evaluates its own generated manuscript) and cross-reviews (a model evaluates manuscripts generated by other models) under identical evaluation conditions. In addition, BERT was used to compute embedding-based semantic similarity as a robustness measure. The results indicated that GPT-4 and GPT-3 exhibited strong self-preference bias, assigning significantly higher scores to their own outputs than to manuscripts generated by other models. Human review scores were more closely aligned with AI cross-reviews than AI self-reviews, whereas LLaMA showed weaker and less consistent self-preference patterns, corresponding to its higher rejection rate during quality screening. These findings remained robust across prompt-sensitivity, resampling, alternative aggregation, subsample, and model-specification checks. Taken together, the findings point to a structural source of distortion in AI-assisted evaluation, suggesting that bias may arise not only from training data but also from how generation and evaluation roles are configured within a workflow. The study contributes to research on AI-enabled evaluation systems by extending current discussions of algorithmic bias toward workflow design, and it offers practical guidance on cross-model evaluation, calibration, and human-in-the-loop safeguards in high-stakes evaluative settings.