AI Research System Publishes 36 AI-Generated Studies Covering LLM Safety and Embodied AI

On October 26, 2026, Chaoyan Intelligence publicly released 36 research results produced by its AI system, spanning AI research and embodied intelligence for robotics. The company said the studies were generated almost entirely by AI with minimal non-expert human intervention, covering the full process from literature review and method design to code implementation, experiment analysis, and paper writing.

In the AI research track, the studies address five problem areas: model evaluation, information extraction, agents, LLM safety and alignment, and model mechanisms. Representative works include research on AI-generated content tracing, safety monitoring of reasoning models, internal mechanisms of vision-language-action (VLA) models, neurosymbolic reasoning, and evaluation reliability. The embodied intelligence track focuses on control and VLA models.

Chaoyan Intelligence said it is releasing the code, experimental data and generation traces to support testable, accumulative AI research that can be continued by the community.