AI-Generated Campaign Ads Impersonate Candidates Without Limits

A campaign advertisement attacking New York Governor Kathy Hochul shows her welcoming undocumented immigrants to luxury hotels, walking through parks filled with homeless people, and appearing to endorse crack dealers outside police precincts. The problem: none of it is real, and the figure on screen isn't actually Hochul—it's an AI-generated deepfake.

The ad, created by Republican challenger Bruce Blakeman's campaign for the November gubernatorial election, carries a small disclosure at the end noting it "includes AI-generated imagery." But the synthetic nature of the content—including fabricated dialogue where the AI version of Hochul makes statements she never made—represents a new frontier in political advertising that operates largely without guardrails.

Why it matters

As generative AI tools become more sophisticated and accessible, political campaigns can now create convincing fake videos of opponents saying or doing things that never happened. Research shows voters struggle to identify AI-generated content, especially when disclosures are easy to miss. Without regulation, campaigns face no meaningful constraints on using this technology to spread disinformation during elections.

Deepfakes spread across political campaigns

Blakeman defended his use of AI to PIX11, stating he uses "every tool available" and comparing AI-generated political content to satire used "since George Washington." His campaign's ad concludes with the slogan "Criminals, Cronies, and Illegal Immigrants for Hochul."

Similar AI-generated content has appeared in races nationwide, according to USA TODAY. In the Los Angeles mayoral race, Republican candidate Spencer Pratt used AI to show women in a Pilates class whispering support for his campaign. In Texas, an AI ad depicted Democratic Senate candidate James Talarico dressed as Maria from "The Sound of Music" singing about transgender issues.

Some AI-generated political content has crossed into explicitly offensive territory. During the 2025 New York City mayoral race, the Andrew Cuomo campaign used AI to depict opponent Zohran Mamdani eating rice with his hands. President Donald Trump previously shared a video using AI to portray Barack and Michelle Obama as apes.

The case for government regulation

The columnist argues that self-regulation by campaigns or technology companies won't address the problem, as both profit from AI-generated political advertising. While Republicans have created more offensive AI content to date, the expectation is that Democrats will eventually adopt similar tactics in response.

The fundamental issue extends beyond partisan politics: AI systems lack any inherent sense of what content is appropriate or ethical to generate. That judgment falls entirely to the politicians commissioning the ads—a group not known for consistent moral restraint during campaign season.

Educating voters to spot deepfakes, while valuable, cannot serve as the primary defense against sophisticated AI-generated disinformation. The only effective solution, according to the analysis, is government regulation of AI use in political advertisements.

These details were first reported by USA TODAY columnist Sara Pequeño.