Disinformation Deepfake Circulates of State Department Spokesman Matthew Miller Suggesting Belgorod Can Be Attacked with U.S. Weapons

A deepfake video of State Department spokesman Matthew Miller falsely suggested Belgorod was a legitimate target for Ukrainian strikes. This disinformation spread on Telegram and Russian media, misleading the public and inciting tensions. U.S. officials condemned the deepfake. This incident is an example of the threat of AI-powered disinformation and hybrid attacks.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.1 Disinformation, surveillance, and influence at scale

The deepfake video was used as part of a coordinated disinformation campaign aimed at manipulating public opinion and political discourse surrounding the war in Ukraine.

Additional risk subdomains

• 3.1 False or misleading information: The AI-generated video fabricated false claims about U.S. foreign policy and the evacuation status of Belgorod, misleading viewers.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by human actors who intentionally created and distributed a deepfake video post-deployment of generative AI technology to spread political disinformation.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3. Limited Risk: The report describes an AI-generated deepfake video. Under the EU AI Act, AI systems that generate or manipulate image, audio or video content (deepfakes) pose a moderate risk and are subject to transparency obligations to ensure users are informed.

AI system and alleged parties

• AI system: None named
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: Unknown deepfake creators
• Alleged deployer: Russian government
• Alleged harmed parties: Matthew Miller, Department of State, Biden administration

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Minor, indirect Severe
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Minor, indirect Minor
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Severe
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation and spread of a deepfake video designed to mislead the public.

Directly caused: A 49-second deepfake video of a U.S. spokesman was created and circulated on Telegram channels and Russian social networks like VK.

Indirectly caused: Russian state media outlets, including TASS, and the Human Rights Council repeated and disseminated the false claims from the video.

Inferred additional harm: The video likely reached millions of viewers across Russia and Ukraine, contributing to a toxic information environment and escalating geopolitical tensions.

Democracy

Reported: The report explicitly describes the video as part of hybrid attacks aimed at eroding governance and democratic systems.

Directly caused: The deepfake video manipulated the likeness of a U.S. official to falsely represent U.S. foreign policy, directly targeting public trust in government communications.

Indirectly caused: The incident prompted international discussions among NATO allies on how to counter foreign state information manipulation and protect democratic systems.

Inferred additional harm: The ongoing use of such deepfakes threatens to permanently degrade public trust in official diplomatic communications and democratic institutions.

Epistemic

Reported: The report explicitly describes the incident as an illustration of the growing threat of AI-powered disinformation.

Directly caused: The video fabricated statements by a U.S. spokesman, falsely claiming that Belgorod was a legitimate target and that other countries allowed deep strikes into Russia.

Indirectly caused: Russian state bodies and media outlets repeated the fabricated remarks verbatim, presenting them as true to the Russian public.

Inferred additional harm: The widespread dissemination of the fake video contributes to the erosion of a shared consensus reality regarding the war in Ukraine.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 340000

Potential causes

Management

• Kremlin Disinformation Strategy: Management of state propaganda used fakes as a core strategy.
• Lack of Platform Moderation: VK and Telegram management failed to flag or remove the deepfake.

Technology

• AI Deepfake Generation Tools: Advanced AI allowed creating realistic fabricated video from actual footage.
• Imperfect Visual Synthesis: Lip sync was off and tie changed color midway, showing technical flaws.

Data Inputs

• Use of Authentic Source Footage: Actual footage of State Dept spokesman was used to seed the deepfake.

Human Factors

• Audience Susceptibility: Belgorod residents and officials believed the fake due to high tension.
• Intentional Malicious Creation: Actors intentionally fabricated remarks to manipulate public opinion.

Process and Methods

• Social Media Distribution Networks: Telegram and VK allowed rapid dissemination of unverified deepfake video.
• Lack of Immediate Fact-Checking: Russian media outlets shared the video without verifying its authenticity.

Regulatory Environment

• Lack of Deepfake Regulations: Absence of international laws banning or labeling state-backed deepfakes.
• State-Controlled Media Ecosystem: Russian state bodies and media amplified the fake without regulatory check.

Information quality

• Classification confidence: High
• Reason for confidence: The report clearly details the creation, content, and dissemination of the deepfake video, as well as the reactions of both U.S. and Russian officials. The role of AI in generating the deepfake is explicitly stated, and the political context is well-documented.
• Ambiguities identified: The specific AI tools or platforms used to create the deepfake are not identified, and the exact origin/creator of the video remains unconfirmed.
• Alternative interpretations: None. The incident is unambiguously a deepfake-based disinformation campaign.

A state-sponsored deepfake video impersonating a U.S. State Department official was disseminated by Russian state media to falsely claim U.S. support for strikes on Russian territory. The incident represents a substantial information warfare threat, highlighting the evolving use of generative AI to fabricate official diplomatic communications and manipulate public opinion during active geopolitical conflicts.

• Overall national security impact: Substantial
• Scope: Multiple nations
• Primary target: United States
• Other affected: Russia, Ukraine, NATO countries
• Alleged perpetrator: Russia

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic capability and disinformation threat rather than an active, unresolved crisis requiring immediate 72-hour tactical response.
• Autonomy: Human-controlled. AI was used as a tool to generate the video, but the creation, distribution, and targeting were entirely directed by human operators.
• Novelty: Evolved capability. Represents an evolution of existing state-sponsored disinformation capabilities, leveraging improved video and voice synthesis to target specific diplomatic figures.

Impact by dimension

• Physical security: Negligible. No physical attacks or critical infrastructure damage resulted from the creation or dissemination of the video.
• Information security: Substantial. State-sponsored deepfake targeting a U.S. official to fabricate foreign policy positions, widely disseminated by Russian state media to manipulate public opinion during an active conflict.
• Sovereignty: Minor. Falsely represented official U.S. diplomatic positions on foreign policy, though it did not disrupt internal government decision-making or constitutional processes.
• Economic security: Negligible. No financial system attacks, critical supply chain disruptions, or strategic technology theft were reported.
• Societal stability: Minor. Caused potential anxiety and distress among residents of Belgorod due to false claims of impending military strikes and evacuations, but did not lead to widespread civil unrest.