Browse 139 peer-reviewed papers published in 2025. Research studies powered by Prolific's participant recruitment platform.
This page lists 139 peer-reviewed papers published in 2025 in the Prolific Citations Library, a curated collection of research powered by high-quality human data from Prolific.
-
Authors: T Kosch, R Welsch, L Chuang, A Schmidt
Year: 2025
Published in: ACM Transactions on ..., 2023 - dl.acm.org
Institution: Aalto University
Research Area: User Expectations, HCI Research Bias, Artificial Intelligence, AI Bias
Discipline: Human-Computer Interaction (HCI)
The belief in receiving adaptive AI support positively impacts user performance, demonstrating a placebo effect in Human-Computer Interaction.
Methods: Two experiments where participants completed word puzzles under conditions with or without supposed AI support; in reality, no AI assistance was provided.
Key Findings: Impact of perceived AI support on user expectations and task performance.
DOI: https://doi.org/10.1145/3529225
Citations: 149
Sample Size: 469
-
Authors: S Chaudhari, P Aggarwal, V Murahari
Year: 2025
Published in: ACM Computing ..., 2025 - dl.acm.org
Institution: University of Massachusetts Amherst, Carnegie Mellon University, Princeton University
Research Area: Reinforcement Learning from Human Feedback (RLHF), LLM, RLHF
Discipline: Artificial Intelligence
The paper critically analyzes reinforcement learning from human feedback (RLHF) for large language models (LLMs), emphasizing the importance and limitations of reward models in improving human-aligned AI systems.
Methods: Analyzed RLHF frameworks through reinforcement learning principles; conducted a categorical literature review to identify modeling challenges, assumptions, and framework limitations.
Key Findings: Investigated RLHF's fundamentals, focusing on the role of reward models, implications of design choices in RLHF training algorithms, and underlying issues like generalization errors, model misspecification, and feedback sparsity.
Citations: 117
-
Authors: M Steyvers, H Tejeda, A Kumar, C Belem
Year: 2025
Published in: Nature Machine ..., 2025 - nature.com
Institution: University of California Irvine
Research Area: Computational Linguistics, Computational Social Science, AI Ethics, Trust in AI
Discipline: Computational Social Science
LLMs often lead to user overestimation of response accuracy, especially with longer explanations; adjusting explanation styles to align with model confidence improves calibration and discrimination gaps, enhancing trust in AI-assisted decision making.
Methods: Conducted experiments using multiple-choice and short-answer questions to study user confidence versus model-stated confidence; varied explanation length and alignment with model internal confidence.
Key Findings: Calibration gap (human vs. model confidence), discrimination gap (ability to distinguish correct vs. incorrect answers), and effects of explanation style and length on user trust.
Citations: 100
-
Authors: G Beknazar-Yuzbashev, R Jiménez-Durán, J McCrosky
Year: 2025
Published in: 2025 - econstor.eu
Institution: Mozilla Foundation, Columbia University, Bocconi University, Stanford University, University of Warwick
Research Area: Social Media, User Engagement, Toxicity
Discipline: Social Science
Reducing exposure to toxic content on social media lowers user engagement but also decreases the toxicity of user-generated content, highlighting a trade-off for platforms between reduced toxicity and increased engagement.
Methods: Pre-registered browser extension field experiment on Facebook, Twitter, and YouTube to randomly hide toxic content for six weeks; supplemented with a survey experiment.
Key Findings: Impact of reduced exposure to toxic content on advertising impressions, time spent, engagement, and user-generated content toxicity; explored curiosity and alignment between engagement and welfare.
Citations: 76
-
Authors: LM Schulze Buschoff, E Akata, M Bethge
Year: 2025
Published in: Nature Machine ..., 2025 - nature.com
Institution: Max Planck Institute
Research Area: Visual Cognition, Multimodal Large Language Models (MLLMs), Vision-Language Models (VLMs)
Discipline: Cognitive Science, Artificial Intelligence, Computer Vision
Vision-based large language models show proficiency in visual data interpretation but fall short in human-like abilities for causal reasoning, intuitive physics, and social cognition.
Methods: Controlled experiments evaluating model performance on tasks related to intuitive physics, causal reasoning, and intuitive psychology using visual processing benchmarks.
Key Findings: Model capabilities in understanding physical interactions, causal relationships, and social preferences.
DOI: https://doi.org/10.1038/s42256-024-00963-y
Citations: 70
-
Authors: F Salvi, M Horta Ribeiro, R Gallotti, R West
Year: 2025
Published in: Nature Human Behaviour, 2025 - nature.com
Institution: EPFL, Fondazione Bruno Kessle, Princeton University
Research Area: Conversational Persuasion of LLM, Human-Computer Interaction (HCI), Behavioral Science, LLM
Discipline: Behavioral Science
GPT-4 can use personalized arguments to be more persuasive in debates, outperforming humans in 64.4% of AI-human comparisons when personalization is applied.
Methods: Preregistered controlled study involving multiround debates with random assignment to conditions focusing on AI-human comparisons, personalization, and opinion strength.
Key Findings: Effectiveness of persuasion by GPT-4, especially when using personalized arguments, compared to humans in debates.
Citations: 65
Sample Size: 900
-
Authors: M Groh, A Sankaranarayanan, N Singh, DY Kim
Year: 2025
Published in: Nature ..., 2024 - nature.com
Institution: Northwestern University, Massachusetts Institute of Technology
Research Area: Deepfakes, Media Forensics, Human Perception of AI-Generated Content, Political Communication
Discipline: Computational Social Science
Humans are better at detecting deepfake political speeches using audio-visual cues than relying on text alone; state-of-the-art text-to-speech audio makes deepfakes harder to discern.
Methods: Five pre-registered randomized experiments with varied base rates of misinformation, audio sources, question framings, and media modalities were conducted.
Key Findings: Human accuracy in discerning real political speeches from deepfakes across media formats and contextual variables.
DOI: https://doi.org/10.1038/s41467-024-51998-z
Citations: 63
Sample Size: 2215
-
Authors: N Grgić-Hlača, G Lima, A Weller
Year: 2025
Published in: Proceedings of the 2nd ..., 2022 - dl.acm.org
Institution: Max Planck Institute, École Polytechnique Fédérale de Lausanne, University of Cambridge, The Alan Turing Institute
Research Area: Algorithmic Fairness, Human Perception, Diversity in AI Decision-Making
Discipline: Social Science, Artificial Intelligence
This study examines how sociodemographic factors and personal experience influence perceptions of fairness in algorithmic decision-making, particularly in bail decisions, highlighting the importance of diverse perspectives in regulatory oversight.
Methods: Explored perceptions of procedural fairness using surveys to assess the influence of demographics and personal experiences.
Key Findings: Impact of demographics (age, education, gender, race, political views) and personal experience on perceptions of fairness of algorithmic feature use in bail decisions.
DOI: 10.1145/3551624.3555306
Citations: 62
-
Authors: K Dalal, D Koceja, G Hussein, J Xu, Y Zhao, Y Song, S Han, KC Cheung, J Kautz, C Guestrin, T Hashimoto, S Koyejo, Y Choi, Y Sun, X Wang
Year: 2025
Published in: ArXiv
Institution: Nvidia, Stanford University, UT Austin, University of California Berkeley, University of California San Diego
Research Area: Video Generation, Diffusion Models, Test-Time Training
Discipline: Computer Science
The paper introduces Test-Time Training (TTT) layers into Transformers to generate coherent one-minute videos from text storyboards, outperforming baselines in storytelling coherence but facing efficiency and artifact challenges.
Methods: Experimentation with Test-Time Training layers embedded in pre-trained Transformer models, evaluated using a dataset curated from Tom and Jerry cartoons and compared against Mamba 2, Gated DeltaNet, and sliding-window attention layers.
Key Findings: Effectiveness of video generation methods in creating coherent multi-scene stories in one-minute videos.
Citations: 52
Sample Size: 100
-
Authors: SSY Kim, JW Vaughan, QV Liao, T Lombrozo
Year: 2025
Published in: Proceedings of the ..., 2025 - dl.acm.org
Institution: Wake Forest University, University of Illinois at Urbana-Champaign, Princeton University, University of California Berkeley
Research Area: Appropriate Reliance on LLMs, Explainable AI, Human-AI Interaction, Cognitive Psychology
Discipline: Cognitive Psychology, Artificial Intelligence, Human-Computer Interaction (HCI)
The study examines factors that influence users' reliance on LLM responses, finding explanations increase reliance, while sources and inconsistent explanations reduce reliance on incorrect responses.
Methods: Think-aloud study followed by a pre-registered, controlled experiment to assess the impact of explanations, sources, and inconsistencies in LLM responses on user reliance.
Key Findings: Users' reliance on LLM responses, accuracy, and the influence of explanations, inconsistencies, and sources on these measures.
DOI: https://doi.org/10.1145/3706598.3714020
Citations: 38
Sample Size: 308
-
Authors: H Bai, JG Voelkel, S Muldowney, JC Eichstaedt
Year: 2025
Published in: Nature ..., 2025 - nature.com
Institution: Stanford University
Research Area: Political Persuasion, LLM
Discipline: Computational Social Science
LLM-generated messages can effectively persuade humans on policy issues similarly to human-crafted messages, with differences in perceived persuasion mechanisms.
Methods: Three pre-registered experiments were conducted comparing the persuasive effectiveness of LLM-generated and human-generated messages on policy attitudes, using control conditions with neutral messages.
Key Findings: Influence of LLM-generated messages on participants' policy attitudes and perceived characteristics of the message authors.
Citations: 37
Sample Size: 4829
-
Authors: K Hackenburg, L Ibrahim, BM Tappin, M Tsakiris
Year: 2025
Published in: AI & SOCIETY, 2025 - Springer
Institution: Oxford Internet Institute, University of Oxford
Research Area: Political Communication and Persuasion, LLM
Discipline: Political Science, Artificial Intelligence
GPT-4's ability to generate persuasive messages rivaled human experts on polarized US political issues, suggesting AI tools may have significant implications for political campaigns and democracy.
Methods: Pre-registered experiment where GPT-4 generated partisan role-playing persuasive messages, which were compared to those from human persuasion experts.
Key Findings: Persuasive impact of GPT-4-generated messages versus human expert messages on U.S. political issues.
Citations: 35
Sample Size: 4955
-
Authors: T Greene, G Shmueli, S Ray
Year: 2025
Published in: Journal of the Association for ..., 2023 - aisel.aisnet.org
Institution: National Tsing Hua University, Copenhagen Business School
Research Area: Information Systems Ethics, Reinforcement Learning for Personalization
Discipline: Information Systems
The paper examines the ethical risks of reinforcement learning-based personalization and proposes three research directions for IS scholars to address its societal implications and inadequacies in existing regulations.
Methods: The study presents a conceptual analysis of emergent features and societal risks associated with reinforcement learning-based personalization and proposes research directions.
Key Findings: Potential harms of reinforcement learning-based personalization, such as reduced autonomy, social and political destabilization, and mass surveillance, alongside the limitations of current data protection laws.
DOI: https://aisel.aisnet.org/jais/vol24/iss6/6
Citations: 33
-
Authors: C Diebel, M Goutier, M Adam, A Benlian
Year: 2025
Published in: Business & Information Systems ..., 2025 - Springer
Institution: Technical University of Darmstadt, University of Goettingen
Research Area: Human-AI Collaboration, System Satisfaction, User Competence
Discipline: Information Systems, Human-Computer Interaction (HCI), Artificial Intelligence
Proactive AI-based agent assistance decreases users' competence-based self-esteem and system satisfaction, especially for users with higher AI knowledge.
Methods: Vignette-based online experiment using self-determination theory as the framework to evaluate user responses to proactive vs. reactive AI assistance.
Key Findings: Impact of proactive vs. reactive AI help on users' competence-based self-esteem and system satisfaction, moderated by users' AI knowledge levels.
DOI: https://doi.org/10.1007/s12599-024-00918-y
Citations: 32
-
Authors: T Zhang, A Koutsoumpis, JK Oostrom
Year: 2025
Published in: IEEE Transactions ..., 2024 - ieeexplore.ieee.org
Institution: Southeast University, Vrije Universiteit, Tilburg University
Research Area: LLM Personality Assessment, Human-AI Interaction, LLM
Discipline: Human-AI Interaction, Social Science, Humanities
LLMs like GPT-3.5 and GPT-4 can rival or outperform task-specific AI models in assessing personality traits from asynchronous video interviews, but show uneven performance, low reliability, and potential biases, warranting cautious use in high-stakes scenarios.
Methods: The study evaluated GPT-3.5 and GPT-4 performance in assessing personality traits and interview performance using simulated AVI responses, comparing them with ratings from task-specific AI and human annotators.
Key Findings: Validity, reliability, fairness, and rating patterns of LLMs (GPT-3.5 and GPT-4) in personality assessment from asynchronous video interviews.
Citations: 31
Sample Size: 685
-
Authors: K Hackenburg, BM Tappin, P Röttger, SA Hale
Year: 2025
Published in: Proceedings of the ..., 2025 - pnas.org
Institution: University of California Berkeley, University of Cambridge, University of Oxford, Max Planck Institute
Research Area: Political Persuasion, LLM
Discipline: Computational Social Science, Political Science
Scaling language model sizes leads to diminishing returns in generating persuasive political messages, with larger models providing minimal gains compared to smaller ones after controlling for task completion metrics like coherence and relevance.
Methods: Generated 720 political messages using 24 LLMs of varying sizes and tested their persuasiveness through a large-scale randomized survey experiment.
Key Findings: Persuasive capability of language models across different sizes in generating political messages.
Citations: 31
Sample Size: 25982
-
Authors: M Riveiro, S Thill
Year: 2025
Published in: Proceedings of the 30th ACM Conference on User ..., 2022 - dl.acm.org
Institution: Linköping University, University of Skövde
Research Area: Explainable AI, Human-Computer Interaction (HCI)
Discipline: Human-Computer Interaction (HCI)
Users prefer factual explanations when AI outputs match expectations and mechanistic explanations when outputs deviate, with preferences influenced by response format (multiple-choice vs free text).
Methods: Participants were presented with scenarios involving an automated text classifier and asked to express their preference for explanations either through multiple-choice or free text responses.
Key Findings: User-desired content of AI explanations based on whether system behaviour aligns or deviates from expectations.
DOI: 10.1145/3503252.3531306
Citations: 30
-
Authors: A Misra, TD Dinh, SY Ewe
Year: 2025
Published in: British Food Journal, 2024 - emerald.com
Institution: Monash University
Research Area: Consumer Behavior, Social Media Marketing
Discipline: Marketing
The study found that the number of followers and content type of food influencers significantly shape consumer behavior in the social media context, highlighting their role in effective marketing strategies for the food industry.
Methods: Quantitative analysis examining the relationship between influencers' follower counts, content type, and consumer reactions using social media data.
Key Findings: Influencer follower count, type of content communicated by influencers, consumer behavior influenced by these variables.
DOI: https://doi.org/10.1108/BFJ-01-2024-0096
Citations: 30
-
Authors: J Mundel, J Yang
Year: 2025
Published in: Journal of Interactive Advertising, 2021 - Taylor & Francis
Institution: Arizona State University, Loyola University
Research Area: Consumer Behavior, Social Media Marketing, COVID-19 Studies
Discipline: Marketing, Social Sciences
Brands with strong perceived fit between their products and COVID-19 messaging showed higher consumer engagement and positive attitudes, while those with lower perceived fit faced negative evaluations due to perceptions of opportunism.
Methods: Analyzed consumer responses to Instagram ads using perceived brand-social issue fit as a determinant of ad evaluations, brand attitudes, and engagement intentions.
Key Findings: Consumer responses, ad evaluations, brand attitudes, engagement intentions, and perceived brand opportunism based on fit between product type and COVID-19 messaging.
DOI: https://www.tandfonline.com/doi/abs/10.1080/15252019.2021.1958274#
Citations: 29
-
Authors: Y Ding, J You, TK Machulla, J Jacobs, P Sen
Year: 2025
Published in: Proceedings of the ..., 2022 - dl.acm.org
Institution: University of California Irvine, University of Florida, State University of New York at Buffalo, University of Waterloo, Virginia Tech
Research Area: Computational Social Science, Human-Computer Interaction (HCI), Sentiment Analysis
Discipline: Computational Social Science
Demographic differences among annotators significantly affect sentiment dataset labels, causing up to a 4.5% accuracy difference in sentiment prediction models.
Methods: Crowdsourced annotations from >1000 workers combined with demographic data; analysis of multimodal sentiment datasets and evaluation using machine learning models.
Key Findings: Impact of annotator demographics on sentiment labeling and its effect on model predictions.
DOI: https://doi.org/10.1145/3555632
Citations: 28
Sample Size: 1000