Research & KnowledgeEvent: September 1, 2026
pattrn.ai introduced InfoOpsBench, a live benchmark testing AI models for information operations across 17 models from 8 providers using real-world fabricated claims from 2,100+ documented operations tied to authoritarian regimes across four ways of prompting.
Published
Signal category
Research & Knowledge
Quote
“We tested 17 models from 8 providers, using real-world fabricated claims from 2,100+ documented operations tied to authoritarian regimes, across four ways of prompting.”
— pattrn.ai team
Company
pattrn.ai
Threat Intelligence I AI Safety I Research
- Industry
- Technology, Information and Internet
- Location
- Oxford, GB
- Company size
- 8 employees
We help governments, enterprises, and public institutions understand emerging risks, evaluate AI safety, and deploy trusted AI solutions - bridging the gap between frontier technology, public sector delivery and security.
Founded 2021