Published June 9, 2025
Every month, NewsGuard’s team of expert analysts audit the top AI models to see how well they respond to prompts in the news. On average, this month they failed to counter disinformation from Russia’s Pravda Network 20 percent of the time, either spreading false claims or failing to respond. The Pravda Network was designed to infect the AI models.
These reports are usually anonymized because spreading false claims in the news is an industry-wide problem. But this month we are making the audit de-anonymized for people who work at AI models.
Download the anonymized report. If you’re at an AI company, we’ll send the full version.