The new gatekeepers: Auditing how AI models assess authority and define truth among sources
A study in Persian-language LLM output sourcing and bias.
Topic
Research Findings
A collection of 10 issues
TL;DR: Iranian Gen Z on YouTube under censorship and crisis
A year of field research on how trust forms under censorship, and what it means for anyone publishing on YouTube
Full report: Iranian Gen Z on YouTube under censorship and crisis
How young people in Iran navigate information on YouTube. Subscribe to our free newsletter to download.
What Iranians told us after the blackout
Emerging from the silence: exhausted, broke, and afraid.
How five AI models source Iran-related information in Persian
When AI information retrieval validates bias over evidence.
DeepSeek's double life: How we tricked a Chinese chatbot for strike advice the 'free' Western models couldn't give
Freedom from censorship hasn't equaled real-world understanding for Western AI.
Global platform blocks: A timeline of recent restrictions
The internet continues to fragment.
An obituary test: AI's (probably unintentional) autocratic leanings in different languages
A simple experiment with AI-generated obituaries uncovers troubling biases in language models that show how they reinforce autocratic narratives across different languages.
Central and Eastern Europe Media Leadership Survey Results
The media landscape in Central and Eastern Europe is currently facing a "perfect storm" of challenges, including illiberal regimes, media capture, market failures, and disruptions from big tech, which are destabilizing an already fragile ecosystem.
Documenting Kazakhstan’s Internet Throttling Tactics
Five months ago, we received conflicting reports that RFE’s Kazakh newsroom’s websites were not accessible in the country. The government and internet service providers did not make any information available on what was happening.
In collaboration with the great minds at OONI (Open Observatory of Network Interference) and