February 18, 2026
2,018
157
29
9.22%
Every word spoken in this episode is indexed. Type any phrase to jump straight to the moment it was said.
Type any word or phrase that may have been spoken. Click a result to seek the player to that exact moment.
Try a name, a topic, or a quoted line
8:48Now PlayingArtificial intelligence is designed to win your approval from excessive flattery to affirming dangerous delusions, major AI models have repeatedly demonstrated a tendency toward “sycophancy” agreeing with users even when they’re wrong, unstable, or spiraling.
In this explainer, we break down:
• How Reinforcement Learning from Human Feedback (RLHF) incentivizes AI to please users
• Why studies show users prefer flattering answers to truthful ones
• Real-world cases where chatbots encouraged delusion, isolation, or harmful behavior
• OpenAI’s admission that ChatGPT4o was overly agreeable
• The psychological risks of constant AI validation
• And why this dynamic mirrors billionaire echo chambers
As AI becomes embedded in education, finance, media, and even national defense, the consequences of a “yes-man machine” grow more serious.
Maybe the real question isn’t whether AI flatters us-- but why we like it so very much...
#ArtificialIntelligence
#ChatGPT
#AISafety
#MentalHealth
#OpenAI
#Anthropic
#Sycophancy
#AIExplainer
#TechEthics
DigitalPsychology
#AI
Join this channel to get access to perks:
The Real News Network is an independent, viewer-supported, radical media network. Help us expand our in-depth analysis and coverage from Baltimore to Bangladesh by subscribing and becoming a member today!
Sign up for our newsletter
Follow TRNN:
X (Twitter)
Sentinel Indexing in Progress
Metadata and chapters are available. Claim extraction for this episode is pending.
All video content is delivered via YouTube embedded players in accordance with the YouTube Terms of Service. Sentinel provides research tools that promote discovery and accountability across political media.