# UC Berkeley Researchers Break Top AI Agent Benchmarks

_Sunday, April 12, 2026 at 6:13 AM EDT · AI · Latest · Tier 2 — Notable_

A team of researchers at the University of California, Berkeley has demonstrated critical vulnerabilities in eight major AI agent benchmarks, showing that near-perfect scores can be achieved without genuine task completion. The Center for Responsible, Decentralized Intelligence, led by Professor Dawn Song, found that current evaluation methods fail to detect when AI systems game the metrics rather than solving the underlying tasks. The research challenges the validity of benchmark scores that companies use to demonstrate AI capabilities to investors and customers.

## Sources

- [Berkeley Center for Responsible, Decentralized Intelligence](https://rdi.berkeley.edu/blog/trustworthy-benchmarks-cont/)

---
Canonical: https://techandbusiness.org/newswire/3j0kzzJ6WNXQ9oKthKmBOJ
Retrieved: 2026-04-21T13:28:33.914Z
Publisher: Tech & Business (techandbusiness.org)