This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A Russian-aligned disinformation network known as Storm-1516 orchestrated a campaign to smear Democratic vice-presidential nominee Tim Walz by circulating a fabricated video. The video featured an individual impersonating a former student, Matthew Metro, and falsely accusing Walz of sexual assault. While some experts debated whether the video was a 'deepfake' or a 'cheap fake' (a real person impersonating another), U.S. intelligence confirmed the content was part of a coordinated Russian effort to interfere in the 2024 U.S. election. The video reached millions of views on social media platforms before being flagged and removed.
A purported viral deepfake accused Democratic vice-presidential nominee Tim Walz of misconduct by using the stolen identity of former student Matthew Metro. Circulated on X and other platforms, the video reportedly reached millions before being flagged for manipulation. U.S. intelligence later revealed it and three other similar events were part of the Russian disinformation campaign Storm-1516, whose aim is to disrupt the 2024 elections.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to conduct large-scale disinformation campaigns, malicious surveillance, or targeted and sophisticated automated censorship and propaganda, with the aim of manipulating political processes, public opinion, and behavior.
Also classified as
The perpetrators stole and exploited the real-world identity of Matthew Metro, impersonating him to bring false accusations against a public figure.
The incident involves a coordinated, state-aligned foreign influence campaign using fabricated media to manipulate public opinion and political processes at scale.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was driven by human actors (Russian state-backed trolls) who intentionally created and spread manipulated media to interfere in the U.S. election.
Added categories peak at 4: Severe (Epistemic Harm, Inferred)
A Russian-aligned propaganda network (Storm-1516) conducted a coordinated disinformation campaign targeting the 2024 US presidential election by distributing a fabricated video smearing vice-presidential nominee Tim Walz. The incident represents a notable foreign influence operation using AI-manipulated media to undermine democratic processes and target political leadership.
The AI systems (such as deepfake generators and audio manipulation tools) were used as passive instruments by human actors to generate and edit deceptive media, acting strictly on direct inputs without autonomous planning.
Limited Risk: The report describes the use of AI to generate deepfakes and manipulated media, which falls under Risk Level 3 due to transparency obligations for AI-generated content.
The primary technology discussed is the generation of a deepfake video depicting a fake whistleblower to spread political disinformation.
There is high consensus among U.S. intelligence, Microsoft, independent researchers, and journalists that the video was a fabricated smear campaign orchestrated by the Russian group Storm-1516. While there is some technical debate among experts on whether the video is a generative AI deepfake or a cheapfake impersonation with audio manipulation, the overall context of the disinformation campaign and its impact are well-documented.
Some might interpret the video as a simple human impersonation (cheap fake) rather than an AI safety incident, though multiple sources and detectors flagged AI involvement in the audio or video generation.
Technology
AI Deepfake Video Generation
Russian actors used AI to create a fake video smearing Tim Walz.
Audio Manipulation Tools
AI tools transformed voice pitch and tone to mimic a real student.
AI-Generated Fake News Sites
Networks used AI to auto-generate articles on spoofed news outlets.
Data Inputs
Open-Source Identity Harvesting
Public profiles were scraped to steal Matthew Metro's biographical data.
Inconsistent Metadata in Emails
Manipulated email screenshots contained formatting and timestamp errors.
Fake Media Verification Inputs
Forged documents like certificates and visas were used as source inputs.
Human Factors
Influencer Amplification
Right-wing influencers shared the video without verifying its authenticity.
Voter Susceptibility to Disinfo
Electorate polarization made users eager to believe damaging rumors.
Identity Theft Exploitation
The victim's real-life details were exploited to build false credibility.
Process and Methods
Inadequate Content Moderation
Social media platforms failed to quickly remove or flag the fake video.
Narrative Laundering Campaigns
Storm-1516 systematically laundered false claims through fake whistleblowers.
Delayed Platform Response Times
Slow response let the video gain over five million views in 24 hours.
Management
Reduced Moderation Investment
Social media firms cut funding for trust and safety operations.
Prioritization of Viral Engagement
Platform algorithms prioritized sensational, highly engaging content.
Inadequate Risk Assessments
Platforms failed to prevent foreign state-sponsored influence campaigns.
Regulatory Environment
Lack of Deepfake Regulations
Absence of strong federal AI laws allowed unchecked deepfake distribution.
Platform Immunity Protections
Constitutional free speech and legal protections shield host platforms.
Jurisdictional Enforcement Limits
US authorities could not easily prosecute actors residing in Moscow.
