How human–AI feedback loops alter human perceptual, emotional and social judgements

2024/12/18 by Moshe Glickman, Tali Sharot · 1 voice · 19 citations
Decision Sciences · Neuroscience · Social Sciences · #Decision-Making and Behavioral Economics #Ethics and Social Impacts of AI #Psychology of Moral and Emotional Judgment

paper · pdf · doi:10.1038/s41562-024-02077-2

openalex publication_date 2024/12/18 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/03

Abstract

Artificial intelligence (AI) technologies are rapidly advancing, enhancing human capabilities across various fields spanning from finance to medicine. Despite their numerous advantages, AI systems can exhibit biased judgements in domains ranging from perception to emotion. Here, in a series of experiments (n = 1,401 participants), we reveal a feedback loop where human-AI interactions alter processes underlying human perceptual, emotional and social judgements, subsequently amplifying biases in humans. This amplification is significantly greater than that observed in interactions between humans, due to both the tendency of AI systems to amplify biases and the way humans perceive AI systems. Participants are often unaware of the extent of the AI's influence, rendering them more susceptible to it. These findings uncover a mechanism wherein AI systems amplify biases, which are further internalized by humans, triggering a snowball effect where small errors in judgement escalate into much larger ones.

Citations

• On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications.
• Bias Amplification in Artificial Intelligence Systems
• Deep Residual Learning for Image Recognition
• Shortcut learning in deep neural networks
• A Systematic Study of Bias Amplification
• Large language models propagate race-based medicine
• Studying and improving reasoning in humans and machines
• Algorithm appreciation: People prefer algorithmic to human judgment
• In AI we trust? Perceptions about automated decision-making by artificial intelligence
• Algorithm aversion: People erroneously avoid algorithms after seeing them err.
• High-Resolution Image Synthesis with Latent Diffusion Models
• ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness

Cited by

• AI-Assisted Programming Decreases the Productivity of Experienced Developers by Increasing the Technical Debt and Maintenance Burden
• A Methodology for Transparent Logic-Based Classification Using a Multi-Task Convolutional Tsetlin Machine
• Embracing Dialectic Intersubjectivity: Coordination of Differential Perspectives in Content Analysis With LLM Persona Simulation
• Taxonomy of User Needs and Actions
• Understanding Mode Switching in Human-AI Collaboration: Behavioral Insights and Predictive Modeling
• Human-AI Narrative Synthesis to Foster Shared Understanding in Civic Decision-Making
• The renaissance of management: When regular employees become multi-level managers of AI agents
• Risks and Realities of Speculative Ethics: Lessons from Nanotechnology for the Artificial Intelligence Discourse
• AI as moral cover: How algorithmic bias exploits psychological mechanisms to perpetuate social inequality
• Why sycophantic LLMs may imperil interactive norms between humans
• How ensembling AI and public managers improves decision-making
• Bias in the Loop: How Humans Evaluate AI-Generated Suggestions
• Going with the Mainstream: Exploring GPT Representation of Journalistic Culture
• The Anti-Ouroboros Effect: Emergent Resilience in Large Language Models from Recursive Selective Feedback
• What do librarians look like? Stereotyping of a profession by generative Ai
• AI as Political Infrastructure: A Diamond Model of Political AI Ethics
• Exploring Human-AI Complementarity in CPS Diagnosis Using Unimodal and Multimodal BERT Models
• Unveiling trust in AI: the interplay of antecedents, consequences, and cultural dynamics
• Visually grounded emotion regulation via diffusion models and user-driven reappraisal

Discussions

• Glickman & Sharot "reveal a feedback loop where human–AI interactions alter processes underlying human perceptual, emotional and social judgements, subsequently amplifying biases in humans." doi.org/1 [bsky, 3 points, 1 comments]

Related

• BIASeD: Bringing Irrationality into Automated System Design
• Rolling in the deep of cognitive and AI biases
• Beyond Isolation: Towards an Interactionist Perspective on Human Cognitive Bias and AI Bias
• Language Models Embody and Amplify Human Cognitive Distortions: What Is to Be Done?
• Humans inherit artificial intelligence biases