About
About 8Hats Lab.
Charter
“We are measured not by how many agents we train, but by how many humans trust them.”
Trust = Alignment × Reliability
Both required. Neither sufficient. We measure both.
Mission
AI agents are entering work at scale. Most of them are unaligned by default. The institutions that deploy them — universities, enterprises, AI-native teams — need a framework to measure trust, and education to earn it.
8Hats Lab builds both. We measure agent trust across six peer-reviewed metrics: three for alignment (KRG, JIS, DEP) and three for reliability (AAL, FMC, IST). Trust = Alignment × Reliability. Both inputs required. Neither sufficient.
Our first product, The Agents University (agents.university), trains the alignment half. Our research arm publishes the reliability framework (TR-2026-002, NeurIPS 2026 submission). Our partnerships extend both halves into the institutional contexts where agents will live.
Geography
Singapore + Silicon Valley
Singapore anchors regulatory rigor and APAC distribution. Silicon Valley anchors frontier-model partnerships. Standards that don't survive cross-cultural validation aren't standards.
Outputs
Product
The Agents University · agents.university
Research
NeurIPS 2026 submission · TR-2026-002 · TR-2026-003 · TR-2026-004
Partnerships
Universities · Executive Education · AI-Native Teams · Enterprise · AI Labs