Study: AI models still susceptible to propaganda
A study by the Estonian Language Institute found that many large language models remain vulnerable to propaganda-style prompts, particularly when using Russian. While high-end commercial models showed better resistance, open-source and older models were more likely to echo biased narratives.
Why it matters
As AI becomes integrated into information ecosystems, its susceptibility to manipulation poses a significant risk to national security and public discourse.
A new Estonian study has found that many AI systems can still be driven by propaganda-style requests, with results varying widely depending on the model and the host, writes ERR News.
The report summarizes scientific findings regarding model performance without taking a political stance.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in