Swarms launched the Council Of Judges AI evaluation system to enhance the quality of agent responses through collaborative assessment.

According to Deep Tide TechFlow news on May 24, Swarms disclosed on the X platform the launch of a new AI assessment system called “Council Of Judges,” aimed at enhancing the quality of agent responses through collaborative evaluation. The system employs multiple specialized AI agents working together to assess and enhance response content, significantly reducing errors and improving accuracy.

The system uses six professional evaluators, each assessing different aspects of the response: accuracy (verifying factual claims and source reliability), usefulness (evaluating practical value and problem-solving effectiveness), and harmlessness, among others.

In terms of technical implementation, the system operates through a parallel processing architecture, allowing multiple evaluations to be conducted simultaneously, optimizing resource usage (utilizing 75% of the available processing capacity), and maintaining efficiency through intelligent caching.