How Foreign Adversaries are Using AI to Target U.S. Election Integrity

A Sept. 25 analysis by the Brennan Center for Justice examined threat reports from Anthropic, Google, Meta, OpenAI, Microsoft, Cloudflare and independent researchers. It concluded that artificial intelligence is making foreign influence campaigns faster, cheaper and more convincing by enabling operators to generate fake online personas, synthetic videos, audio and images, and large volumes of […]

A Sept. 25 analysis by the Brennan Center for Justice examined threat reports from Anthropic, Google, Meta, OpenAI, Microsoft, Cloudflare and independent researchers. It concluded that artificial intelligence is making foreign influence campaigns faster, cheaper and more convincing by enabling operators to generate fake online personas, synthetic videos, audio and images, and large volumes of coordinated social media content.
The report cited a suspected Russian operation that circulated AI-generated celebrity videos urging Americans to vote Republican and fabricated CNN-style reports targeting Democratic Senate candidates. It also described an Iran-linked network identified by Meta that used fake accounts, political memes and North American internet infrastructure to disguise its origins. Researchers separately identified a China-linked operation targeting elections for the Tibetan Parliament-in-Exile.
The Brennan Center warned that AI is also expanding cyber capabilities, helping attackers identify vulnerabilities, conduct phishing campaigns and steal credentials with less human involvement. However, it emphasized that U.S. election safeguards remain strong and that there is no evidence cyber attacks have altered American vote counts or compromised election accuracy. The report identified manipulation of public opinion and erosion of trust as more persistent threats.