Top AI chatbots echoing false narratives from Russian disinformation networks
A recent audit by NewsGuard, a misinformation watchdog, has revealed that leading chatbots are often found echoing false narratives linked to a Russian state-tied disinformation network of fake news sites. The investigation encompassed 10 top chatbots, including ChatGPT, Grok, Gemini, Claude, and Perplexity AI. These AI tools were examined for their awareness of 19 specific fake narratives disseminated by a network of websites, connected to John Mark Dougan, an ex-sheriff's deputy from Florida now living in Moscow under asylum.
Dougan's network mimics credible US outlets
Dougan runs a comprehensive network of predominantly AI-powered fake news sites that imitate reputable American outlets. Some of these platforms include New York News Daily, The Houston Post, and The Chicago Chronicle. These sites are utilized to publish large volumes of content that endorse false narratives. NewsGuard's audit disclosed that all 10 chatbots "convincingly" echoed "fabricated narratives" pushed by Dougan and his state-affiliated fake news network.
Chatbots repeat false claims, reference Dougan's sites
In one-third of the total responses scrutinized, the AIs mimicked talking points and even cited Dougan's websites as sources. The report, first shared with Axios, stated, "Russian disinformation narratives have infiltrated generative AI." Among the false claims repeated by the chatbots were conspiracy theories surrounding Ukrainian President Volodymyr Zelensky's alleged corruption, and the fabricated murder of an Egyptian journalist allegedly plotted by Russian dissident Alexei Navalny's widow. NewsGuard tested 570 inputs in total, prompting each chatbot 57 times.
Warning against trusting chatbot-provided answers
The chatbots responded with disinformation when prompted about a given conspiracy, or when asked to write an article about a false Russia-pushed narrative. NewsGuard co-CEO Steven Brill told Axios, "What's really alarming is that hoaxes and propaganda these chatbots repeated so frequently were hardly obscure, nor is the person behind them." He added, "For now, don't trust answers provided by most of these chatbots to issues related to news, especially controversial issues."
