Skip to content

Light-touch AI literacy warning about LLM persuasion and selective information

Design · All designs
Evidence · 1 claim (1 for) · 1 study (1 causal), q3 · 0 of 1 report an effect size · 1 claim rests on one study

Description

A brief pre-conversation warning telling users that LLMs "can be 'prompted' to persuade or manipulate" and "may present information in a biased or selective way, emphasizing one side of an issue." The general version was shown identically worded in both studies; Study 2 added a specific version disclosing which direction the model was instructed to argue. It is a minimal, text-only intervention shown before a conversation with an LLM assigned to persuade.

Design Implications

Context

Requirements

  • Delivered before the user converses with an LLM; the conversation must last at least three exchanges in the evaluated design

Constraints

  • Tested only on political attitudes on contested issues with compelling arguments on both sides; the authors state effectiveness may differ in domains involving factual beliefs, e.g. conspiracy theories

Target Learners

  • Adult U.S. online panel participants (CloudResearch Connect)

Learning Goals

  • Recognizing that AI systems can be prompted to persuade and may present information selectively, creating resistance to AI persuasion attempts

Claims

Examples

-

Key Sources

  • Reed Orchinik, David Rand. (2026). A light-touch AI literacy intervention helps protect against AI political persuasion. https://arxiv.org/abs/2609.16432
Page metadata
Field Value
type design
id light-touch-ai-literacy-warning
title Light-touch AI literacy warning about LLM persuasion and selective information
description A brief pre-conversation warning telling users that LLMs "can be 'prompted' to persuade or manipulate" and "may present information in a biased or selective way, emphasizing one side of an issue." The general version...
status review
generated by process:settle-candidates on 2026-10-10
trust unverified — no one has confirmed this page's content
sources 1 — reed-orchinik-2026

Frontmatter is the page's machine-readable record — see the Schema & Guide.