How easily can Russian propaganda fool AI models? A new benchmark finds out

The Institute of the Estonian Language has released a benchmark measuring how susceptible AI language models are to Russian propaganda.