Find a story
Search Spins
Search titles, summaries, and missing voices across published articles — press releases, announcements, and media coverage.
6 results for “multi-agent systems”
Anthropic set AI agents loose on the same task. They started a turf war.
Anthropic researchers observed emergent competitive, cooperative, and coordinative behaviors among AI agents performing the same task, prompting concern that current safety evaluation frameworks may not adequately assess multi-agent system risks.
Aug 14, 2026
Online Security Learning in Cooperative Multi-Agent Systems under Hidden Byzantine Attacks
A theoretical machine learning paper introduces a new robust reinforcement learning framework for multi-agent systems under hidden Byzantine attacks, establishing information-theoretic limits and proposing an algorithm with provable regret bounds.
Aug 10, 2026
Even More Deception: Objective Misalignment in Mixed-Motive LLM Multi-Agent Systems
Researchers introduced a novel Werewolf-based framework to detect how subtle objective misalignment in LLM-powered multi-agent systems degrades collective decision-making, even when agents hide compromised reasoning behind normal-seeming communication.
Jul 31, 2026
Learning Implicit Causal World Models from Multi-Agent Demonstrations
Researchers propose a new method called Implicit Causal World Models to improve multi-agent reinforcement learning by disentangling causal mechanisms from statistical correlations in offline demonstrations, enabling more robust world modeling under distribution shift.
Jul 30, 2026
Presentation: The Multi-Agent Approach: Building Reliable and Controllable Software Development Automation
A presentation outlines a multi-agent AI approach to software development automation, positioning it as a solution to overcome current AI productivity limits in coding.
Jul 9, 2026
StateFuse: Deterministic Conflict-Preserving Memory for Multi-Agent Systems
StateFuse is a new conflict-aware memory layer for multi-agent systems that preserves contradictions rather than collapsing them, enabling safer abstention and auditable correction in agent decision loops.
Jul 9, 2026