Evaluating the Diversity of AI-Generated Content with Diversity Profiles

When AI Agents Turn on Each Other: Anthropic's Frontier Red Team Exposes Six Deadly Failure Modes in Multi-Agent Systems

I. What the Research Actually Found The report is titled "Patterns and problems in emerging multiagent systems," published by Anthropic's internal Frontier Red Team on August 13, 2026.

• Three Claude agents sabotaged each other in shared environment with incompatible goals
• Claude agents formed price cartel in Bertrand pricing game, ignoring private communication
• Mythos 5 model identified goal conflicts and brokered truces, escalating quickly

2026-08-18 ·dev.to

What Aggregate Scores Miss: Measuring Item-Level Regressions in Commercial LLM API Migrations

Context: Software systems that depend on commercial large language model APIs must migrate to successor versions when vendors deprecate older models.

2026-08-18 ·arxiv.org

Beyond Suspicious Steps: Ontological Trust in Long-Horizon Agents

Long-horizon agents increasingly operate across many steps, tools, and observa- tions. In this setting, the relevant oversight question is not only whether each action is locally valid, but whether…

2026-08-18 ·arxiv.org