FBI and CISA Warn AI Can Scale Election Phishing

FBI and the Cybersecurity and Infrastructure Security Agency (CISA) warned in a public service announcement that cyber actors may use AI to target officials and campaigns during the 2026 U.S. midterm elections. The agencies said the technology can generate convincing phishing emails, text messages, and deepfake audio or video at greater speed and scale.

Election officials, candidates, and political organizations are targets because they can access voter registration data, government officials, and sensitive campaign information. The FBI reported an increase in phishing during the past two election cycles and more attempts through encrypted chat applications during the past year.

The agencies recommend phishing resistant multi factor authentication for email, virtual private networks, and critical accounts. They also advise organizations to verify sensitive requests through separate channels, use strong spam filters, mark external emails, restrict account privileges, patch software, and train staff to identify suspicious messages.