Biased AI improves human performance but reduces perceived helpfulness
cs.HC, cs.AI, cs.CY
Submitted: 2025-08-12
Updated: 2026-09-18
Code: https://github.com/ShiyangLai/Biased_AI_Enhance_DM
License: http://creativecommons.org/licenses/by/4.0/
The gist: Artificial intelligence (AI) increasingly shapes how people think, engage, and evaluate information.
Terminology
Abstract
Artificial intelligence (AI) increasingly shapes how people think, engage, and evaluate information. To minimize risk, most current systems are designed to present as ideologically neutral with standardized output. Yet growing evidence suggests that these principles suppress cognitive engagement, impair human decision-making, and erode societal diversity. Here we test the opposite approach by deliberately injecting bias into AI assistants. In three randomized experiments with 5,000 participants, biased AI improved human performance relative to default and neutral AI in tasks ranging from misinformation evaluation and financial investment to graduate education. These gains carried a subjective cost. Participants systematically undervalued AI they believed to be biased and inflated the helpfulness of AI they believed to be neutral, regardless of the systems' actual behavior. Interacting with two AIs whose biases flanked the participant's own perspective preserved the performance gains while limiting the subjective cost and one-sided influence. Our findings reveal the strategic value of intentional bias in AI design. Rather than performing a single fair, reliable, and authoritative voice, AI that speaks from specific viewpoints triggers cognitive agency and elevates human-AI performance in judgment, decision-making, and problem-solving.
Sources
- Representation Engineering: A Top-Down Approach to AI Transparency
- Constitutional AI: Harmlessness from AI Feedback
- A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity
- Position: Towards Bidirectional Human-AI Alignment
- Four Bottomless Errors and the Collapse of Statistical Fairness
- Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task
- Towards Understanding Sycophancy in Language Models
- Multi-Agents are Social Groups: Investigating Social Influence of Multiple Agents in Human-Agent Interactions
- Collaborating with AI Agents: Field Experiments on Teamwork, Productivity, and Performance
- AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation
- Steering Large Language Models using Conceptors: Improving Addition-Based Activation Engineering
- Refusal in LLMs is an Affine Function
- Steering Language Models With Activation Engineering
- Prompting Techniques for Reducing Social Bias in LLMs through System 1 and System 2 Cognitive Processes
- Political DEBATE: Efficient Zero-shot and Few-shot Classifiers for Political Text
- From Pretraining Data to Language Models to Downstream Tasks: Tracking the Trails of Political Biases Leading to Unfair NLP Models
Related papers
- EduGage: A Multimodal Dataset and Benchmark for Sensor-Based Momentary Assessment of Engagement in Self-Guided Video Learning
- EvoDesign: Agentic Editable Diagram Creation via Design Expertise Evolution
- HAGI++: Head-Assisted Gaze Imputation and Generation
- Linking Behaviour and Perception to Evaluate Meaningful Human Control over Partially Automated Driving
- Review of Explainable Decision Support and Adaptive Human-Machine Interfaces for Automation Transparency in Maritime Autonomous Surface Ships
- Towards Cognitive Process-Aware Proactive Writing Support