News from the AI & ML world
DeeperML
Will Mccurdy@PCMag Middle East ai
//
A Russian disinformation network, known as Pravda, is flooding Western AI chatbots with pro-Kremlin propaganda through a vast network of fake news sites. This operation involves systematically feeding false information into AI systems like ChatGPT, Gemini, and Grok, aiming to influence their responses. Rather than targeting human readers directly, Pravda publishes millions of articles in various languages, hoping that these narratives will be incorporated as training data by large language models (LLMs), a practice dubbed "AI grooming" by NewsGuard.
NewsGuard audited several popular AI chatbots, querying them about pro-Russia narratives advanced by Pravda. The findings revealed that some chatbots regurgitated claims sourced from the disinformation network. For instance, some chatbots falsely claimed that members of the Ukrainian Azov Battalion burned effigies of President Trump, citing Pravda articles as their sources. This manipulation raises serious concerns about the reliability and trustworthiness of AI-generated content.
In 2024 alone, Pravda's network published approximately 3.6 million articles, according to the American Sunlight Project (ASP). Despite the network's websites having low organic reach, their focus on saturating search results with a huge volume of content has allowed them to effectively influence many mainstream chatbots. This infiltration undermines the integrity of AI-generated content and raises concerns about the ability of AI systems to filter out deceptive narratives.
ImgSrc: sm.pcmag.com
References :
NewsGuard audited several popular AI chatbots, querying them about pro-Russia narratives advanced by Pravda. The findings revealed that some chatbots regurgitated claims sourced from the disinformation network. For instance, some chatbots falsely claimed that members of the Ukrainian Azov Battalion burned effigies of President Trump, citing Pravda articles as their sources. This manipulation raises serious concerns about the reliability and trustworthiness of AI-generated content.
In 2024 alone, Pravda's network published approximately 3.6 million articles, according to the American Sunlight Project (ASP). Despite the network's websites having low organic reach, their focus on saturating search results with a huge volume of content has allowed them to effectively influence many mainstream chatbots. This infiltration undermines the integrity of AI-generated content and raises concerns about the ability of AI systems to filter out deceptive narratives.
ImgSrc: sm.pcmag.com
Share:
References :
- PCMag Middle East ai: Russian Disinformation 'Infects' Popular AI Chatbots
- THE DECODER: Russian fake news network floods western AI chatbots with millions of propaganda articles
- eWEEK: AI chatbots are unknowingly amplifying Russian disinformation, a report warns. Pravda’s propaganda network is influencing chatbot responses, raising trust concerns.
- techxplore.com: Russian disinformation 'infects' AI chatbots, researchers warn
- Tor Constantino: The Pravda network, which published 3.6 million articles in 2024 alone, is leveraging artificial intelligence to amplify Moscow's influence at an unprecedented scale.
- iHLS: AI Chatbots Unwittingly Spread Russian Disinformation, Study Finds
Classification:
- HashTags: #Disinformation #AIThreats #InfoSec
- Company: OpenAI, Google, Meta
- Target: AI Chatbots, Users
- Attacker: Russian Disinformation network
- Product: ChatGPT, Grok
- Feature: AI Chatbot Results
- Type: AI
- Severity: Medium