What's Changed Since Our Last Report

Our previous coverage tracked the broader California primary landscape — turnout numbers, polling, prediction markets. Now the focus has shifted to a specific and accelerating battleground: AI-generated political advertising is hitting races at every level simultaneously, and the controversies are stacking up fast.

Three separate incidents — one in Santa Barbara County, one in Los Angeles, one in Texas with direct national implications — have broken in quick succession. AI-generated attack ads are no longer a novelty. They're a pattern.

Santa Barbara: The Clown Ad That Blew Up

The Santa Barbara County Republican Party texted voters an AI-generated ad depicting Democrat-endorsed county supervisor candidate Ricardo Valencia as a literal clown — red nose, curly wig, colorful suit — wandering through scenes of homelessness, fire, and crime in Santa Maria.

The voiceover used what critics called a stereotypical Latino accent. One scene showed the AI Valencia in clown makeup, phone in hand for a selfie, as a man robbed an elderly woman near a Shell gas station nearby.

Bobbi McGinnis, chair of the Santa Barbara County Republican Party, told the Santa Barbara News-Press the organization was "simply experimenting with AI content" and denied the party is racist.

That defense landed badly. California Senate President Monique Limon, joined by two other state lawmakers, released a joint statement: "There is no place in our community for the use of racist caricatures, disinformation, or deceptive AI-generated content designed to inflame division and demean candidates based on their background or identity."

County Supervisor Laura Capps called it "vile and unacceptable." Santa Maria Councilwoman Gloria Soto called it "a racist caricature" and "disgraceful."

Valencia, a high school teacher and Santa Maria-Bonita School Board member, called the ad "racist and dehumanizing" in an Instagram story, according to the NY Post.

Using a stereotypical accent overlaid on a Latino candidate's AI likeness is difficult to defend. Critics of Valencia's politics could make their case without it. The Santa Barbara GOP handed their opponents a potent talking point and made their actual arguments — about crime and homelessness — harder to hear.

Los Angeles: Batman, Luke Skywalker, and Spencer Pratt

In the L.A. mayor's race, AI content is operating on a completely different register — and it's fan-made, not party-made.

Reality TV personality Spencer Pratt's supporters have produced a stream of hyper-cinematic AI videos: Pratt as Batman storming City Hall, Pratt as Luke Skywalker on an Imperial speeder bike as Governor Gavin Newsom plays Emperor Palpatine and Mayor Karen Bass plays Darth Vader.

The Los Angeles Times reported on May 18, 2026 that these videos are "heavily shared" and generating real buzz — but political experts interviewed for the piece are skeptical they translate into actual votes.

Hany Farid, a digital forensics professor at the University of California, Berkeley, noted about the NRSC's Texas deepfake (more on that below) that AI fakes are becoming "hyper-realistic" to the point where most viewers would NOT immediately identify them as fake.

NBC News reported May 7, 2026 that the Pratt AI ads are "being celebrated by some, but also raise ethical questions" — particularly as California primary debates loom. Dana Griffin's NBC report flagged the timing directly: candidates are heading into high-stakes debates while AI content about them proliferates unchecked.

Texas: The National GOP Goes Full Deepfake

The most technically significant development comes from outside California but sets the template for what's coming here.

On March 11, 2026, the National Republican Senatorial Committee posted an 85-second AI deepfake of James Talarico, the Democratic nominee in the Texas U.S. Senate race. A fabricated, lifelike version of Talarico appears to speak directly to the camera for over a minute — reading real tweets from 2021 about transgender issues, race, and religion, interspersed with fake self-praising commentary the real Talarico never said.

CNN reported on March 13, 2026 that Farid reviewed the video and called it "hyper-realistic," adding he does NOT think most people would immediately know it was fake. An "AI GENERATED" disclaimer appeared on screen — but CNN noted it was small, faint, and confined to a bottom corner.

The NRSC's position, per a source familiar with their thinking cited by CNN: AI is a "consistently effective" way to highlight opposing candidates' real statements.

That argument works when applied to actual quotes. It breaks down when the same video puts fabricated words into the candidate's mouth. There's a clear distinction between "showing what someone actually said" and "making them say things they never said." The NRSC crossed that line.

What Mainstream Coverage Is Getting Wrong

Left-leaning outlets are treating this almost entirely as a Republican problem. It's not. The L.A. mayor's race AI content is largely fan-generated and ideologically diffuse. The technology is party-agnostic. Democrats will use it too — and some already are.

Right-leaning coverage has largely ignored the Santa Barbara story, possibly because the clown-accent ad is hard to defend on the merits.

Neither side is seriously engaging with the core legal vacuum: California has passed some AI disclosure laws, but enforcement is weak, platform guardrails are described by the L.A. Times as inadequate, and First Amendment challenges have stalled federal action, according to CNN's reporting.

What Voters Face Now

If you see a political video between now and the June primary, you have no reliable way to know if the candidate actually said what you're watching them say. The disclosure requirements that exist are too small, too easy to miss, and frequently ignored.