A user posted an actual Monet Water Lilies painting labeled “Made with AI” and invited critics to explain its inferiority; confident, detailed takedowns followed.
Key Takeaways
X user @SHL0MS used the platform’s “Made with AI” label on a real Monet from the 250-painting Water Lilies series to bait art critics.
Critics produced detailed technical complaints about spatial coherence, color saturation, depth, and composition – all fabricated flaws in a master work.
One commenter wrote an 850-word critique; many participants deleted replies once the deception went viral.
A 2024 Nature study found people prefer AI art over human art when unlabeled, but rate it lower once they know the source – independent of actual quality.
The 2004 Kruger effort heuristic study supports this: perceived effort drives perceived value, not visual inspection.
Hacker News Comment Review
Commenters noted the irony doubles: the viral “850-word breakdown” cited in the article was itself likely AI-generated, adding another layer of fabricated authority.
The thread landed on a broader point: art perception is primarily a cultural and social game, not an optical one – inspection of the work alone cannot settle the dispute.
Consensus leaned toward epistemic humility: the experiment shows people are primed to hallucinate flaws when given a framing prompt, not that they lack visual skill.
Notable Comments
@saaaaaam: Flags that the celebrated 850-word critique reads as AI-authored itself – “Fair warning before I dig in” is a tell.
@mayliu2000: “We didn’t get worse at judging art. We just got better at doubting everything.”
