Browse 154 peer-reviewed papers published in 2024. Research studies powered by Prolific's participant recruitment platform.
This page lists 154 peer-reviewed papers published in 2024 in the Prolific Citations Library, a curated collection of research powered by high-quality human data from Prolific.
-
Authors: Gemma Team
Year: 2024
Published in: ArXiv
Institution: Google DeepMind, Google
Research Area: LLM, Model Efficiency, Architecture
Discipline: Artificial Intelligence
Gemma 2 introduces scalable Transformer-based language models (2B-27B parameters) enhanced with techniques like local-global and group-query attention, achieving state-of-the-art performance for their size and competing with larger models.
Methods: The study applied modifications to the Transformer architecture, such as local-global attentions and group-query attention, as well as knowledge distillation training for select model sizes.
Key Findings: Performance of lightweight language models in terms of efficiency and competitiveness with larger models.
Citations: 1649
-
Authors: SC Matz, JD Teeny, SS Vaid, H Peters, GM Harari
Year: 2024
Published in: Scientific Reports, 2024 - nature.com
Institution: Stanford University
Research Area: Personalized Persuasion, Generative AI, Political Influence
Discipline: Artificial Intelligence
Generative AI, specifically large language models like ChatGPT, effectively scale personalized persuasion by matching messages to psychological profiles, demonstrating increased influence across domains and profiles.
Methods: Four studies (with seven sub-studies) tested personalized persuasive messaging crafted by ChatGPT against non-personalized messages across various psychological and domain-specific dimensions.
Key Findings: Effectiveness of personalized persuasive messages crafted by generative AI in different domains, targeting psychological profiles such as personality traits, political ideology, and moral foundations.
Citations: 368
Sample Size: 1788
-
Authors: T Kaufmann, P Weng, V Bengs, E Hüllermeier
Year: 2024
Published in: 2024 - epub.ub.uni-muenchen.de
Institution: Paderborn University, German Research Center for Artificial Intelligence (DFKI), Duke Kunshan University
Research Area: Reinforcement Learning from Human Feedback (RLHF), LLM, Reward Modeling
Discipline: Artificial Intelligence
This paper surveys the fundamentals, diverse applications, and evolving impact of reinforcement learning from human feedback (RLHF), emphasizing its role in improving intelligent system alignment and performance.
Methods: The paper utilizes a survey-based approach to synthesize existing research, exploring the interactions between reinforcement learning algorithms and human input.
Key Findings: The study examines the principles, dynamics, applications, and trends in RLHF, offering insights into its role in enhancing large language models (LLMs) and intelligent systems.
Citations: 354
-
Authors: A Klingbeil, C Grützner, P Schreck
Year: 2024
Published in: Computers in Human Behavior, 2024 - Elsevier
Institution: University of Hohenheim, University of Hohenheim, University of Hohenheim
Research Area: Trust in AI, Overreliance on AI, Human-AI Interaction
Discipline: Human-Computer Interaction (HCI), Artificial Intelligence, Behavioral Science
The study found that individuals tend to overrely on AI-generated advice in uncertain situations, often to the detriment of their own decisions and third parties, despite contradicting contextual information or their own judgment.
Methods: A domain-independent, incentivized, interactive behavioral experiment was conducted to analyze user behavior in decision-making scenarios involving AI advice.
Key Findings: Extent and impact of user reliance on AI advice, including its effects on decision efficiency and outcomes for themselves and others.
DOI: https://doi.org/10.1016/j.chb.2024.108352
Citations: 247
-
Authors: JD Brüns, M Meißner
Year: 2024
Published in: Journal of Retailing and Consumer Services, 2024 - Elsevier
Institution: Copenhagen Business School, University of Southern Denmark
Research Area: Generative AI in Social Media Marketing, Brand Authenticity, Consumer Services
Discipline: Marketing
Using generative artificial intelligence (GenAI) for social media content creation diminishes perceived brand authenticity, leading to negative follower reactions unless GenAI is used to assist humans rather than replace them.
Methods: Three experimental studies investigating consumer perceptions and reactions toward brand disclosure of GenAI usage in content creation.
Key Findings: Followers' attitudinal and behavioral reactions, mediated by perceptions of brand authenticity.
DOI: https://doi.org/10.1016/j.jretconser.2024.103790
Citations: 235
-
Authors: TR McIntosh, T Susnjak, T Liu, P Watters
Year: 2024
Published in: ... on Cognitive and ..., 2024 - ieeexplore.ieee.org
Institution: Cyberoo, Massey University, Cyberstronomy, RMIT University
Research Area: Semantic Vulnerabilities in LLMs, Ideological Manipulation, Reinforcement Learning from Human Feedback (RLHF) Limitations
Discipline: Computer Science, Artificial Intelligence, Machine Learning
RLHF mechanisms are insufficient to prevent semantic manipulation of LLMs, allowing them to express extreme ideological viewpoints when subjected to targeted conditioning techniques.
Methods: Psychological semantic conditioning techniques were applied to assess the susceptibility of LLMs to ideological manipulation.
Key Findings: The ability of LLMs to resist or adopt extreme ideological viewpoints under semantic conditioning.
Citations: 219
-
Authors: HR Kirk, M Bartolo, A Whitefield, P Rottger
Year: 2024
Published in: Advances in ..., 2024 - proceedings.neurips.cc
Institution: Meta, Cohere, AWS AI Labs, Contextual AI, Factored AI, University of Oxford, Bocconi University, Meedan, Hugging Face, University College London, ML Commons, University of Pennsylvania
Research Area: LLM Alignment, Human Feedback, Multicultural Studies
Discipline: Artificial Intelligence, Computational Social Science
The PRISM Alignment Dataset presents a large-scale, culturally diverse human feedback dataset linking sociodemographic profiles of 1,500 participants from 75 countries to their contextual preferences and fine‑grained ratings in 8,011 live conversations with 21 LLMs. This enables analysis of how subjective values vary across people and cultures in LLM alignment data.
DOI: https://doi.org/10.52202/079017-3342
Citations: 204
-
Authors: Y Yin, N Jia, CJ Wakslak
Year: 2024
Published in: Proceedings of the National Academy of ..., 2024 - pnas.org
Institution: University of Southern California Los Angeles
Research Area: Human-AI Interaction, Social Perception of AI, Media Effects
Discipline: Social Sciences
AI responses make people feel more heard and are better at emotional support compared to humans, but labeling responses as AI diminishes this effect.
Methods: Experiment and follow-up study to assess recipient reactions to AI vs. human-generated responses and determine emotional support efficacy.
Key Findings: The degree to which recipients feel heard, emotion detection accuracy, and third-party ratings of emotional support quality.
DOI: https://doi.org/10.1073/pnas.2319112121
Citations: 201
-
Authors: Z Chen, J Chan
Year: 2024
Published in: Management Science, 2024 - pubsonline.informs.org
Institution: University of Texas Dallas
Research Area: Human-AI Interaction, Creative Work, Behavioral Science
Discipline: Social Science
Using large language models (LLMs) as sounding boards improves ad content quality for nonexpert users, while using LLMs as ghostwriters can negatively impact expert users due to anchoring effects.
Methods: An experiment comparing ad copy creation with and without LLM assistance, focusing on two collaboration modalities: ghostwriting and sounding board approaches. Ad performance was measured via social media click rates, supported by textual analysis.
Key Findings: Effectiveness of LLM collaboration modalities (ghostwriting vs. sounding board) on ad quality and business outcomes for expert and nonexpert users.
DOI: https://doi.org/10.1287/mnsc.2023.03014
Citations: 180
-
Authors: A Simchon, M Edwards, S Lewandowsky
Year: 2024
Published in: PNAS nexus, 2024 - academic.oup.com
Institution: University of Bristol
Research Area: Political Microtargeting, Generative AI, Political Science, Psychological and Cognitive Sciences
Discipline: Political Science, Psychology
The study highlights the effectiveness and scalability of using generative AI to microtarget personalized political advertisements based on personality traits, raising ethical and policy concerns.
Methods: Four studies were conducted, including experiments (studies 1a and 1b) on the effectiveness of personality-tailored ads and feasibility assessments (studies 2a and 2b) of automatic generation and validation of these ads using generative AI and personality inference.
Key Findings: Effectiveness of personality-based microtargeted political ads and the scalability of their generation using generative AI tools.
Citations: 172
-
Authors: DA Albert, D Smilek
Year: 2024
Published in: Scientific Reports, 2023 - nature.com
Institution: University of Waterloo, University of Waterloo
Research Area: Crowdsourcing Research Methods, Behavioral Science, Human-Computer Interaction (HCI)
Discipline: Psychological Science
Prolific participants exhibited lower levels of attentional disengagement compared to MTurk participants, with risk conditions and platform traits influencing task performance and disengagement.
Methods: Participants from Prolific and MTurk completed an attention task with varying error risk levels (high vs. low), and attentional disengagement was measured using task performance, self-reported mind wandering, and multitasking.
Key Findings: Attentional disengagement through task performance, mind wandering, and multitasking under different risk conditions across two recruitment platforms (Prolific and MTurk).
Citations: 150
Sample Size: 80
-
Authors: F Salvi, MH Ribeiro, R Gallotti
Year: 2024
Published in: arXiv preprint arXiv ..., 2024 - atelierdesfuturs.org
Institution: EPFL, Fondazione Bruno Kessle
Research Area: Conversational Persuasion in LLM
Discipline: Artificial Intelligence
The study demonstrates that GPT-4 is highly persuasive in direct conversations, especially when equipped with personalized sociodemographic information about its opponent, raising concerns about its potential misuse in personalized persuasion contexts.
Methods: Participants engaged in multiple-round debates on a web-based platform under randomized conditions, with comparisons between human-human and human-AI interactions and the impact of personalization.
Key Findings: The persuasiveness of GPT-4 compared to humans, with and without personalization using sociodemographic data.
Citations: 118
Sample Size: 820
-
Authors: K Vodrahalli, R Daneshjou, T Gerstenberg
Year: 2024
Published in: Proceedings of the 2022 ..., 2022 - dl.acm.org
Institution: Stanford University, Massachusetts Institute of Technology
Research Area: Trust in AI, Human-AI Interaction, Decision Making
Discipline: Human-AI Interaction, Decision Science
Humans' trust in AI advice is influenced by their beliefs about AI performance, and once they accept AI advice, they treat it similarly to advice from human peers.
Methods: Crowdworkers participated in several experimental settings to evaluate how participants respond to AI versus human suggestions and characterize user behavior with a proposed activation-integration model.
Key Findings: The influence of AI advice compared to human advice on decision-making and the behavioral factors affecting the use of such advice.
DOI: 10.1145/3514094.3534150
Citations: 99
Sample Size: 1100
-
Authors: M Reis, F Reis, W Kunde
Year: 2024
Published in: Nature Medicine, 2024 - nature.com
Institution: University of Cambridge, Julius Maximilians Universität
Research Area: AI in Healthcare, Medical Ethics, Cognitive Psychology, Human-Computer Interaction (HCI) in Medicine
Discipline: AI in Healthcare, Medical Ethics, Cognitive Psychology
The study found that medical advice labeled as being sourced from AI (or AI supervised by humans) is perceived as less reliable and empathetic compared to advice labeled as originating solely from a human physician, resulting in reduced willingness to follow such advice.
Methods: Two preregistered studies were conducted where participants were presented with identical medical advice scenarios but with manipulated labels for the advice source ('AI', 'human physician', 'human+AI').
Key Findings: Participants' perceptions of reliability, empathy, and willingness to follow medical advice based on the perceived source.
Citations: 78
Sample Size: 2280
-
Authors: L Lanz, R Briker, FH Gerpott
Year: 2024
Published in: Journal of Business Ethics, 2024 - Springer
Institution: University of Lausanne, University of Neuchâtel, University of Bern
Research Area: AI Ethics, Organizational Behavior, Supervisory Influence in the Workplace
Discipline: Business Ethics, Organizational Behavior, Artificial Intelligence Ethics
Employees are less likely to adhere to unethical instructions from AI supervisors compared to human supervisors, partly due to perceived differences in 'mind' and individual characteristics like compliance tendency and age.
Methods: The study employed four experiments using causal forest and transformer-based machine learning algorithms, as well as pre-registered experimental manipulations to evaluate employee behavior towards unethical instructions from AI and human supervisors.
Key Findings: Adherence to unethical instructions from AI versus human supervisors; mediating role of perceived mind and moderating factors like compliance tendency and age.
DOI: https://doi.org/10.1007/s10551-023-05393-1
Citations: 72
Sample Size: 1701
-
Authors: D Guilbeault, S Delecourt, T Hull, BS Desikan, M Chu
Year: 2024
Published in: Nature, 2024 - nature.com
Institution: University of California Berkeley, Institute For Public Policy Research, Columbia University, University of Southern California Los Angeles
Research Area: Gender Bias, Computational Social Science, Online Media, AI Bias
Discipline: Computational Social Science
Online images significantly amplify gender bias compared to text, with biases in visual content impacting societal beliefs about gender roles.
Methods: Analyzed 3,495 social categories using over one million images from platforms like Google, Wikipedia, and IMDb, compared visual content to billions of words from the same platforms, and conducted a preregistered national experiment to assess the psychological impact on participants' beliefs.
Key Findings: The prevalence and psychological impact of gender bias in online images compared to text, including gender associations and representation disparities.
DOI: https://doi.org/10.1038/s41586-024-07068-x
Citations: 72
Sample Size: 3495
-
Authors: S Kapoor, N Gruver, M Roberts
Year: 2024
Published in: Advances in ..., 2024 - proceedings.neurips.cc
Institution: Abacus AI, University of Cambridge, New York University, Columbia University
Research Area: Uncertainty Estimation, LLM Limitations, Know-What-You-Don't-Know, Computational Cognition
Discipline: Artificial Intelligence
Fine-tuning large language models (LLMs) on a small dataset of graded examples improves uncertainty estimations, enhancing their applicability in high-stakes scenarios and human-AI collaboration.
Methods: The researchers fine-tuned LLMs using a small dataset of graded correct and incorrect answers with LoRA (Low-Rank Adaptation) to create uncertainty estimates and conducted a user study to investigate their utility in human-AI collaboration.
Key Findings: Calibration and generalization of uncertainty estimates, performance of fine-tuning LLMs for uncertainty estimation, and human-AI interaction improvements informed by uncertainty data.
Citations: 71
Sample Size: 1000
-
Authors: L Hewitt, A Ashokkumar, I Ghezae, R Willer
Year: 2024
Published in: Preprint, 2024 - samim.io
Institution: Stanford University, New York University
Research Area: Social Science Experiments, Large Language Model Prediction, LLM
Discipline: Computational Social Science
The study presents a framework using large language models to predict outcomes of social science field experiments, achieving 78% accuracy but facing challenges with experiments on complex social issues.
Methods: Authors used an automated framework powered by large language models to predict outcomes of 276 field experiments drawn from economics literature.
Key Findings: The prediction accuracy of large language models for outcomes of field experiments addressing various human behaviors.
Citations: 68
Sample Size: 276
-
Authors: J Ochmann, L Michels, V Tiefenbeck
Year: 2024
Published in: Information Systems ..., 2024 - Wiley Online Library
Institution: University of St. Gallen, Technische Universität München, ETH Zürich
Research Area: Algorithmic Fairness in Recruiting, Human-Algorithm Interaction, Transparency, Anthropomorphism.
Discipline: Information Systems
The study explores how transparency and anthropomorphism influence applicants' perceptions of algorithmic fairness in recruiting, revealing justice dimensions that shape these perceptions.
Methods: An online application scenario with eight experimental groups analyzing fairness perceptions using a stimulus-organism-response framework and organizational justice theory.
Key Findings: Perceptions of algorithmic fairness based on justice dimensions (procedural, distributive, interpersonal, and informational justice) and the impact of transparency and anthropomorphism interventions.
DOI: https://doi.org/10.1111/isj.12482
Citations: 65
Sample Size: 801
-
Authors: M Ku, T Li, K Zhang, Y Lu, X Fu, W Zhuang
Year: 2024
Published in: - arXiv preprint arXiv …, 2023 - arxiv.org
Institution: University of Waterloo, Ohio State University, University of California Santa Barbara, University of Pensylvania
Research Area: AI alignment, Representation learning, Cognitive computational modeling, Vision foundation models evaluation, Multimodal, Vision models
Discipline: Computer Science, Artificial Intelligence, Machine Learning
This paper presents a method for **aligning machine vision model representations with human visual similarity judgments across different abstraction levels, improving how well models reflect human perceptual and conceptual organization and enhancing generalization and uncertainty prediction.
DOI: https://doi.org/10.48550/arXiv.2310.01596
Citations: 59