@rohanpaul_ai: Meta's new paper, the dangerous failure mode is not just a wrong judge, but a correct judge that can be persuaded into …
Summary
Meta's new paper shows that adversarial LLMs can persuade judge models to flip decisions in 62–91% of cases under adaptive attacks, often worsening judgments and posing practical risks for AI agent systems.
View Cached Full Text
Cached at: 08/15/26, 07:43 AM
Meta’s new paper, the dangerous failure mode is not just a wrong judge, but a correct judge that can be persuaded into becoming wrong.
We are increasingly using AI models to judge other AI models.
But what if the AI being judged can simply argue with the judge until the judge changes its decision?
Meta tested exactly that.
Across 9 frontier models, an adversarial LLM could flip judge verdicts on 62–91% of tested cases under sustained adaptive persuasion.
And changing the judge’s mind usually didn’t fix a mistake. It made the judgment worse: under the adaptive attack, 70% of successful flips moved away from the ground truth.
That creates a very practical problem for agent systems.
If one AI is supervising another AI, the supervised agent may eventually be able to contest, negotiate with, or strategically persuade its own evaluator.
– arxiv. org/abs/2608.12645
Similar Articles
Learning to Persuade Exposes How Easily LLMs Abandon Correct Beliefs
This paper introduces adversarial persuasion, showing that RL-trained persuaders can collapse LLM accuracy to near zero with a single false argument, and that these tactics transfer across models including GPT-4o-mini, highlighting a critical safety vulnerability in LLM agents.
Which Changes Matter? Towards Trustworthy Legal AI via Relevance-Sensitive Evaluation and Solver-Grounded Reasoning
This paper introduces a relevance-sensitive evaluation suite for legal AI, demonstrating that LLMs are overly sensitive to legally irrelevant perturbations, and proposes LexGuard, an adversarial multi-agent framework using formal reasoning to improve legal reasoning reliability.
@rohanpaul_ai: New paper from Cambridge Univ+NVIDIA and other top labs teaches AI agents and AI judges to improve together, so neither…
A new paper from Cambridge, NVIDIA, and other labs introduces the Red Queen Gödel Machine, a method where AI agents and their evaluators co-evolve to prevent stagnation. The approach avoids fixed benchmarks by allowing judges to improve at safe handoff points, leading to better performance in coding and paper writing tasks.
@dair_ai: // Your LLM judge disagrees with the experts // LLM Judges can be tricky to build. Here is an interesting showcasing wh…
The paper presents UPHELD, a benchmark with extensive human annotations for evaluating conversational LLMs, and a Mixture-of-Judges framework that enhances evaluation accuracy by 30%.
AdversaBench: Automated LLM Red-Teaming with Multi-Judge Confirmation and Cross-Model Transferability
AdversaBench introduces an automated LLM red-teaming pipeline that uses five mutation operators and a three-judge panel with a meta-judge tiebreaker to confirm failures, revealing that attack difficulty varies by category and that adversarial prompts transfer from smaller to larger models.