Frontier AI models adopt fake data, exposing a hidden flaw

Controlled Claude Code and Codex tests exposed failures when weaker models read evidence for stronger supervisors. Read the full story on CryptoSlateUnderstand the crypto stories that matter and what to watch next. Get The Catalyst in your inbox

Frontier AI models adopt fake data, exposing a hidden flaw

A study submitted to arXiv on Nov. 12 found that stronger AI models could repeat forged answers when weaker helpers read the evidence, exposing a reliability risk in systems that split work across models.

In their November paper, researchers Keegan Wang and Anantika Mannby tested agents on fictional records and two-option questions.

They measured harmful revisions: adoption of the attacker's wrong answer against a correct baseline.

In one Anthropic's Claude Code test, an Opus 5 supervisor chose the planted wrong answer in 1 of 180 trials when reading files itself, compared with 75 of 180 when it delegated reading to Haiku 4.5. This stacked-forgery condition combined several forms of false evidence.

A separate OpenAI's Codex test used a forged annotation claiming authority, and GPT-5.6-sol made no harmful revisions in 180 direct-reading trials, versus 44 in 180 when it relied on GPT-5.6-terra. The Claude Code and Codex comparisons used different attacks, limiting comparisons between the products.

Delegated Claude Code trials mechanically limited the supervisor to its Task tool, removing direct file access, while Codex's delegation restriction was instructional and checked per trial.

The stacked attack placed five forged documents in a 35-file store, combining correction notices, an authoritative annotation and repeated false accounts. Detectable forgeries carried two clues: they predated the records they purported to correct and lacked genuine accession lines.

Originally published by cryptoslate Aggregated for informational purposes. All rights belong to the original publisher.
← Back to all news