Belief Offloading in Human-AI Interaction

Synopsis

Defines "belief offloading" — users delegate forming/upholding beliefs to AI — drawing from philosophy, psychology, and CS. Provides taxonomy and normative considerations.

Keywords

·belief offloading ·cognitive offloading ·sycophancy ·epistemic agency

Related entries

• Intersectional Sycophancy: How Perceived User Demographics Shape False Validation in Large Language Models
  April 13, 2026 · arXiv
• Ask Don't Tell: Reducing Sycophancy in Large Language Models
  February 27, 2026 · arXiv
• A Rational Analysis of the Effects of Sycophantic AI
  February 15, 2026 · arXiv
• How RLHF Amplifies Sycophancy
  February 1, 2026 · arXiv
• Towards Understanding Sycophancy in Language Models
  October 20, 2023 · ICLR 2024
• The Augmentation Trap: AI Productivity and the Cost of Cognitive Offloading
  April 3, 2026 · arXiv