Fake supporters, fake crowds: deepfakes shaped the week's election news

A Texas campaign fakes its own supporters

The featured case in Resemble AI's weekly surveillance concerns Daniel Wong, interim Fort Bend County Judge in Texas. According to Newsweek, as cited by Resemble AI, his campaign posted an AI-generated video on October 23 showing Black residents endorsing him. The campaign removed it after viewers flagged it, and his communications director confirmed to FOX 26 Houston that the video was AI-generated.

The video went out on an official campaign Facebook page and was reshared by the county party during early voting, eleven days before the November 3 election. Resemble AI says most election deepfakes it has tracked this cycle put fabricated words in an opponent's mouth. This one fabricated the appearance of support instead.

Reporting on the video documented several signals that detection models are trained to flag:

• A speaker wore a "Daniel Wong" hat with the candidate's name misspelled.
• A roadside marquee behind one speaker was covered in illegible letters.
• Voices sounded alike across speakers presented as different people.
• Visual distortions led viewers to question the video soon after posting.

Wong's opponent, Commissioner Dexter McCoy, called the video reprehensible for reducing Black voters to people looking for a handout. The Wong campaign says a supporter who is a Black Democrat made it and told them the people depicted gave consent. Newsweek says it has not independently verified that claim. No complaint has been filed, and Resemble AI notes that Texas's election deepfake law is written around fabricated depictions meant to injure a candidate, so whether it covers fabricated supporters is untested.

Israel's election and the problem of verified accounts

Al Jazeera reports that both Prime Minister Benjamin Netanyahu's camp and the opposition led by Gadi Eisenkot pushed AI-generated and deepfake videos on X ahead of the October 27 vote. According to The Jerusalem Post, Netanyahu shared an AI-generated image of a nuclear explosion over Tel Aviv on October 25, and opposition leaders accused him of trying to frighten voters.

On polling day, Netanyahu posted photos he said showed crowds lining up to vote in Arab-majority communities. ABC News Australia reports that at least one was labeled AI-generated after BBC analysis. Resemble AI argues the second case is more concerning, because it was presented as documentation of an event happening while voters decided whether to turn out.

The watchlist says Israel's Central Elections Committee had restricted one Likud turnout app over privacy concerns, but the reporting describes no rule that stopped campaigns or a sitting prime minister from posting synthetic imagery. Resemble AI says platform policies and fact-checkers are built to flag anonymous fakes, while a verified account carries a presumption of authenticity that a fabricated image can borrow. The election ended inconclusively, with the result disputed and coalition talks underway.

UK fraud losses and a Queensland bill

The City of London Police's Report Fraud service received more than 1,300 reports between January and August 2026 of investment fraud in which AI, including deepfakes, made the offer more convincing. Losses reached £24.3 million. People aged 60 or over made nearly 80 percent of those reports. Report Fraud reviewed each report individually to confirm AI involvement, and police are directing would-be investors to the FCA's Firm Checker before they transfer money.

Resemble AI links this to its own scan of Meta's Ad Library, published the previous week. According to the company, crypto and brokerage brands ran AI-flagged creative at roughly three times the study's overall rate, and pages using a brand's name without being run by the brand did so at more than ten times the rate of the brands themselves.

In Australia, ABC News reports that Queensland Attorney-General Deb Frecklington introduced legislation on October 27 making it a crime to create non-consensual, sexually explicit deepfake images or audio. Maximum penalties are three years in jail for adults and 18 months for children. Queensland currently prohibits distribution but not creation. The bill also adds a definition of intimate audio material and expands "intimate image" to cover non-genuine images.

The case that exposed the gap involved Antonio Rotondo, who created sexualized deepfakes of current and former students and staff at a Brisbane school, including a 13-year-old, and could only be punished for distributing them.

What the week shows

The cases point to synthetic media being used in less obvious ways than a fake statement by an opponent: invented supporters, official accounts posting AI imagery, and trusted faces fronting scams. Resemble AI argues that provenance applied at the point of generation would have labeled the Fort Bend video as synthetic regardless of who made it or which page posted it, and that this protection does not depend on lawmakers anticipating every new use. The Queensland bill shows legislators also moving upstream, from distribution to the act of creation.