Skip to content
Submit incident
AII-1000-53047
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

Russia-Linked Influence Operation Used AI-Cloned Celebrity Voices in 2026 U.S. Midterm Disinformation

Russian government reportedly conducted a covert disinformation campaign using AI-generated deepfake videos of celebrities to target Democratic candidates ahead of the 2026 U.S. midterm elections, according to U.S. intelligence assessments. The operation aimed to undermine confidence in elections and amplify domestic divisions through social media.

10 September 2026 ยท Russian Government ยท Misinformation and synthetic media
10 September 2026
AII-1000-53047View source โ†—
Quick facts
Category
Misinformation and synthetic media
Source
nytimes.com
Date
10 September 2026
Drafted by
Raidu Incident Detection

What happened

The Russian government reportedly authorized a digital influence campaign using artificial intelligence to create fake videos of Hollywood celebrities discussing key battleground Senate races, according to U.S. intelligence assessments.

The operation targeted Americans through social media in hopes of seeding or amplifying domestic divisions around the November 2026 midterm elections.

Reported impact

The campaign appeared aimed at undermining public confidence in the midterm elections and sowing chaos rather than aiding one particular party, though observed examples exclusively attacked Democratic candidates in swing states.

Index analysis

U.S. intelligence officials stated the disinformation campaign appeared less extensive or coordinated than some of Russia's prior election interference efforts, noting Moscow as less interested in midterm races than presidential contests and distracted by its war with Ukraine.

Governance controls

The U.S. intelligence community identified the operations through classified assessments and determined that voting machinery was not targeted in a way that could compromise the actual voting process.

Classification

Category
Misinformation and synthetic media
Industry
Media, social platforms and marketing
Harm status
Potential harm
Intent
Intentional
Timing
Unknown
Harm types
Public interest, Reputational
Harmed
General public, Civil society
Failure modes
Voice clone, Deepfake video, Manipulation or dark pattern
Consequences
None recorded
AI systems
AI Voice Cloning Technology, Deepfake Video Generation

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

nytimes.com
Russia-Linked Influence Operation Used AI-Cloned Celebrity Voices in 2026 U.S. Midterm Disinformation