Tue Jul 08 12:48:13 UTC 2025: Here’s a news article summarizing the provided text:

**Headlines Hysteria? Report Claiming AI Chatbots Amplify Russian Disinformation Under Scrutiny**

**[City, State] –** A recent report claiming that AI chatbots like ChatGPT are readily spreading Russian disinformation is facing heavy criticism, with experts questioning the methodology and conclusions drawn. The initial report, published by NewsGuard, asserted that chatbots repeated false narratives from the pro-Kremlin “Pravda network” a significant 33% of the time. This sparked widespread concern, making headlines in major publications like *The Washington Post* and *Forbes*.

However, independent researchers are casting doubt on these findings. Critics point to NewsGuard’s lack of transparency, as they did not publicly release their “prompts” – the questions used to test the chatbots – making independent verification impossible. Moreover, the study design is accused of being inherently biased, with a focus on specific, often obscure topics related to the Pravda network and the inclusion of responses urging caution as examples of disinformation.

“The study set out to find disinformation – and it did,” said [Name & Affiliation – if provided in original text].

Alternative research, systematically testing major AI chatbots like ChatGPT, Copilot, Gemini, and Grok, found a much lower rate of false claims, approximately 5%. Researchers observed that when disinformation did appear, it was often linked to “data voids” – instances where mainstream outlets lacked coverage of a particular topic, causing the chatbots to pull from less reputable sources. The researchers say this is not necessarily a sign of deliberate Kremlin “grooming” of AI, but rather a consequence of information scarcity.

Experts are warning against overhyping the threat of Kremlin-led AI manipulation, suggesting that such alarmist framings can be counterproductive, leading to repressive policies, eroding trust in credible information sources, and distracting from other, potentially more dangerous applications of AI by malicious actors, such as malware generation. They emphasize the need for careful analysis and a balanced perspective, separating genuine risks from inflated fears in the ongoing conversation about disinformation.

Read More