Scammers clone candidates’ voices to steal cash and voter data ahead of 2026 midterms, FBI warns

A warning has been issued by the FBI regarding fraudulent schemes targeting voters and donors ahead of the 2026 midterm elections, noting that scammers are utilizing advanced technology to deceive the public. These fraudsters are impersonating political candidates and campaigns by cloning voices and creating convincing digital messages. The scams aim to solicit fake campaign donations, sell non-existent merchandise, or harvest personally identifiable information (PII) for subsequent identity theft.
This threat is characterized by the increasing realism provided by AI tools, which complicates detection efforts for victims. Furthermore, scammers are employing investment schemes promising guaranteed returns on contributions, alongside data harvesting tactics that trick individuals into providing voter registration details via fake websites. Concurrently, social media platform X has reported suspending over 13,000 accounts linked to suspected foreign influence operations originating from Iran, Russia, and China targeting U.S. political dialogue.

FBI warns public about financial fraud schemes exploiting the 2026 US Midterm Election.
Scammers are borrowing candidate names, logos, and slogans to collect fake donations and harvest personal data.
Fraudsters use AI tools to create convincing voices, images, and written messages impersonating candidates and government entities.
One scheme involves directing voters to a fake state voter registration page via text or email to steal PII.
Scammers promise victims guaranteed returns on contributions through an investment in a campaign pool.
The FBI reminds the public that campaign donations are not investments and will not be returned.
Social media platform X suspended over 13,000 accounts involved in three suspected foreign influence operations.
These operations were linked to Iran, Russia, and China and targeted US political conversation across the platform.
Cofense's Cyber Intelligence Team Manager advised pausing and verifying messages through official sources like FEC or vote.gov.

The current situation illustrates a rapid technological migration in influence operations, shifting from clumsy social engineering toward sophisticated mimicry powered by generative AI. The strength of this narrative lies in its timely specificity—the impending 2026 election cycle serves as an acute vulnerability window exploited by threat actors seeking financial gain and data harvesting. The core assumption underlying the public safety warnings is that individual vigilance remains a primary, if imperfect, defense against digitally perfected fraud.
This dynamic echoes historical patterns of political opportunism but accelerates them exponentially; where past scams required rudimentary technical skill or persuasive scripting, modern attacks leverage synthetic media to achieve near-perfect credibility at minimal cost. The secondary threat highlighted by X—foreign influence operations—complicates the picture: while domestic actors pursue profit through financial fraud and data theft, state adversaries appear focused on manipulating the information ecosystem itself. This suggests a confluence of financially motivated crime intersecting with geopolitical influence campaigns during sensitive electoral periods.
The primary ethical concern arises from the erosion of epistemic authority; when an AI can perfectly mimic trust—be it a candidate’s voice or a trusted government link—the very anchors of public verification become fungible targets. Who ultimately bears the cost? The individual targeted for data theft, who loses both money and sovereignty over their personal identity.
How does this hyper-realistic landscape change the responsibility dynamic between platform moderation (like X's Civic Integrity flagging) and fundamental digital literacy required of the citizen? If verification sources themselves become potential targets for spoofing, what non-human mechanisms must be developed to restore trust in foundational communication channels?
Patterns detected: none