CIRISAgent open source ethical AI framework for accountable autonomy

We present CIRISAgent, an open-source artificial intelligence framework that
reimagines how autonomous systems interact with humans and each other through
transparent, explainable architecture. Unlike traditional AI systems that
operate as black boxes whose decision-making remains opaque, we build
transparency directly into its structure through a 22-service architecture
organized around clear action verbs and ethical reasoning. As a complete and
deployed agent framework currently deployed in Discord communities, we
demonstrate early promise but require rigorous academic validation. We explore
how we address key challenges in AI alignment through practical engineering
choices, while acknowledging where further research and validation are needed.
See https://ciris.ai and https://github.com/cirisai/cirisagent/ for the ethical
framework and the source code respectively, and https://agents.ciris.ai for the
running agent dashboard.

https://ciris.ai/research-status/ and https://ciris.ai/coherence-ratchet/ have the latest status