This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A NewsGuard audit of 10 leading AI chatbots found that they repeated Russian disinformation narratives in approximately 32% of responses. The models frequently cited fake local news sites as reliable sources, authoritatively presenting false claims as fact. This behavior demonstrates how AI can inadvertently validate and amplify disinformation campaigns, posing risks to public information integrity during election cycles.
An audit by NewsGuard revealed that leading chatbots, including ChatGPT-4, You.com’s Smart Assistant, and others, repeated Russian disinformation narratives in one-third of their responses. These narratives originated from a network of fake news sites created by John Mark Dougan (Incident 701). The audit tested 570 prompts across 10 AI chatbots, showing that AI remains a tool for spreading disinformation despite efforts to prevent misuse.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
Also classified as
The chatbots amplified coordinated foreign influence campaigns designed to manipulate public opinion and political processes during election cycles.
The AI models failed to perform reliably by failing to recognize propaganda websites and repeating known falsehoods.
The chatbots repeated and validated false Russian propaganda narratives, presenting fabricated stories and fake news sources as factual information to users.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by deployed AI chatbots unintentionally generating and repeating Russian disinformation narratives when prompted by users.
Added categories peak at 4: Severe (Epistemic Harm, Inferred)
An audit of ten leading AI chatbots revealed they repeated Russian disinformation narratives in 32% of responses, citing fake local news sites as reliable sources. This highlights how generative AI can be exploited to amplify coordinated foreign propaganda, posing a substantial threat to information security and democratic integrity during global election cycles.
The report describes the AI systems as chatbots that users turn to for quick, customized information and answers to questions, which fits the Assistant behavior type.
Limited Risk: The report describes chatbots and generative AI systems, which are classified as limited risk under the EU AI Act and are subject to transparency obligations to ensure users know they are interacting with an AI.
The systems evaluated are explicitly described as conversational AI chatbots designed to answer user queries and provide information.
The reports provide highly detailed, quantitative results from a structured audit conducted by NewsGuard across 10 specific chatbots. The methodology, prompt counts, and specific false narratives are clearly documented, leaving little ambiguity about the AI systems' behavior.
None
Technology
Inability to Detect AI Personas
Chatbots failed to recognize AI-generated personas like Olesya Movchan.
Vulnerability to Malign Prompts
Models generated fake whistleblower scripts when prompted by bad actors.
Lack of Source Credibility Logic
Chatbots treated fake local news sites as highly credible sources.
Data Inputs
Ingestion of Propaganda Outlets
Models trained on or retrieved data from 167 fake Russian news sites.
Unverified YouTube Testimonies
Chatbots used fake YouTube whistleblower videos as authoritative sources.
Mimicked Trustworthy Domains
Fake sites used names of defunct real newspapers to dupe training filters.
Human Factors
Exploitation by Malicious Actors
Malign actors intentionally prompt chatbots to produce convincing fake news.
User Overreliance on Chatbots
Users increasingly rely on chatbots for quick and customized news updates.
Process and Methods
Inadequate Safeguards for News
AI companies failed to implement special filters for news and election topics.
Ineffective Source Verification
Lack of processes to verify the authenticity of cited web domains.
Weak Guardrails for Election Info
Pledges to curb election misinformation were not backed by robust actions.
Management
Prioritizing Growth over Safety
AI companies deployed models without sufficient safety guardrails for news.
Inadequate Risk Assessment
Failure to assess risks of Russian disinformation networks on election safety.
Unresponsive Safety Teams
AI companies failed to respond to NewsGuard's inquiries regarding findings.
Regulatory Environment
Absence of Clear AI Regulations
Governments are still striving to regulate generative AI safety effectively.
Political Backlash on Watchdogs
Investigations into watchdogs like NewsGuard complicate safety enforcement.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
