journal article Nov 06, 2025

AI hallucinates because it’s trained to fake it till it makes it

View at Publisher Save 10.1126/science.aed6538
Abstract
Teaching chatbots to say “I don’t know” could curb falsehoods but harm business model
Topics

No keywords indexed for this article. Browse by subject →

Metrics
5
Citations
0
References
Details
Published
Nov 06, 2025
Vol/Issue
390(6773)
Pages
558-559
Cite This Article
Celina Zhao (2025). AI hallucinates because it’s trained to fake it till it makes it. Science, 390(6773), 558-559. https://doi.org/10.1126/science.aed6538
Related

You May Also Like

Electric Field Effect in Atomically Thin Carbon Films

K. S. Novoselov, A. K. Geim · 2004

61,289 citations

Optimization by Simulated Annealing

S. Kirkpatrick, C. D. Gelatt · 1983

44,123 citations

Emergence of Scaling in Random Networks

Albert-László Barabási, Réka Albert · 1999

35,859 citations

Judgment under Uncertainty: Heuristics and Biases

Amos Tversky, Daniel Kahneman · 1974

27,432 citations

The Tragedy of the Commons

Garrett Hardin · 1968

22,676 citations