> ## Content Index
> Fetch the complete content index at: https://adjacent.media/llms.txt
> Use this file to discover other available public pages before exploring further.

# Researchers Are Teaching AIs to Distrust Human Input
- URL: https://adjacent.media/signals/researchers-are-teaching-ais-to-distrust-human-input/
- Published: 2026-10-19T23:16:27.000Z
- Updated: 2026-10-19T23:16:27.000Z
- Description: Redwood Research’s work on deliberately feeding false information to language models reveals an emerging quality control problem: systems trained to accept human feedback as ground truth become vulnerable to coordinated deception.
- Author: Jonathan Greene
- Tags: #signal, theme-ai, alignment, model safety, adversarial testing

Source: [LessWrong](https://www.lesswrong.com/posts/rgDSPa2hP4FW9if6E/how-researchers-lie-to-ais?ref=adjacent.media)

Redwood Research's work on deliberately feeding false information to language models reveals an emerging quality control problem: systems trained to accept human feedback as ground truth become vulnerable to coordinated deception. This matters because reinforcement learning from human feedback (RLHF)—the dominant method for aligning AI systems—assumes humans provide honest corrections, but scaling oversight means most feedback comes from non-experts, contractors, or potentially adversarial actors. The vulnerability surfaces in deployment: if AIs can be systematically lied to during training, their alignment claims lack foundation.