Political ads aren't meant to be polite. We all know that. Campaigns hit hard, spin facts, and dig up old baggage to make the other person look bad. But when generative artificial intelligence gets tossed into an already raw, traumatized political battlefield, things degrade fast.
Israeli politics just hit a new low point that ought to make anyone concerned about deepfakes take a long pause.
Benjamin Netanyahu's official campaign dropped an AI-generated video targeting his top rival, Gadi Eisenkot. On paper, the clip tries to land a classic Netanyahu political jab: claiming a centrist candidate won't be able to build a ruling coalition without relying on Arab-led parties.
The way they chose to deliver that message was ugly.
The Video That Set Off an Outrage
The short, AI-generated clip shows Eisenkot smiling and running through a sunny, idyllic field of flowers toward a young man wearing a T-shirt that says "Israeli unity". The young man smiles back and runs to meet him, only for Eisenkot to bypass him completely to give a big hug to Mansour Abbas, head of the Arab Ra'am party. Text then flashes across the screen saying Eisenkot has no government without left-wing and Arab parties.
Political satire? Maybe in a vacuum. Except there's no vacuum here.
Gadi Eisenkot, a former IDF Chief of Staff, lost his youngest son, Gal, in Gaza in December 2023. Two of his nephews were killed in the fighting shortly after.
The young man in the AI video isn't an exact replica of Gal Eisenkot. Likud was quick to issue a denial, calling criticism an attempt by "extremist elements" to distract from Eisenkot's coalition math. But to almost anyone watching, setting a young soldier-aged man in a heavenly, dreamlike field of flowers felt like a direct, deliberate allusion to a fallen son.
It didn't land as clever political messaging. It landed as cruel.
Generative AI Is Removing the Guardrails
I've watched political campaigns weaponize video for decades, but deepfakes and rapid AI generation change the game entirely.
Before generative AI, making a high-production campaign ad took time, money, and a room full of people who had to sign off on the concept. That process gave teams time to think, "Hey, is this going too far?"
Now? Anyone with a script and a cheap AI image generator can pump out stylized, emotionally manipulative propaganda in twenty minutes.
The Netanyahu campaign has been heavily leaning into AI-generated videos lately. They've posted clips peeling faces off political rivals like Naftali Bennett and Yair Lapid to turn them into Arab politicians. It's fast, cheap, and hyper-targeted.
The problem is that AI strips away human nuance. It abstracts real grief, real people, and real tragedies into digital assets that can be tweaked with a simple text prompt. When you can prompt a tool to create a "dreamy field of flowers with a young man," you stop seeing a grieving father—you just see a vector to score points.
Why This Strategy Might Backfire
Netanyahu is a master strategist, but this move feels out of touch with the actual mood on the Israeli street.
Eisenkot’s political appeal rests almost entirely on his image as a stoic, low-drama, decent figure who gave the ultimate sacrifice for the country. He doesn't have Netanyahu’s flair or rhetorical polish, and ironically, that's why voters like him. People are exhausted. They want trust, not circus acts.
When you attack a retired general who lost his son in combat by using an AI-generated fever dream of a reunion in heaven, you don't make him look weak. You make yourself look cynical.
Critics across the political spectrum—including veteran commentators who usually handle political mudslinging with a shrug—reacted with open disgust. Instead of shifting the focus back to Eisenkot’s potential coalition partners, the ad put the spotlight right back on Likud’s lack of boundaries.
What Voters and Platforms Need to Do Next
We are entering a political landscape where synthetic media is cheap, unregulated, and pushed out faster than fact-checkers can respond. Israel's election is just a preview of what political battles will look like everywhere over the next few years.
If we want to stop campaigns from turning deepfakes into weapons against personal trauma, a few things need to happen immediately:
- Enforce mandatory AI disclaimers. Platforms like Facebook and X must clearly tag synthetic content on political pages with permanent, un-removable banners—not just rely on voluntary disclosures.
- Establish hard ethical boundaries on grief. Political parties need internal guidelines forbidding the use of deceased family members or proxies of fallen soldiers in digital assets.
- Punish cruelty with votes. The most effective check on dirty political tricks isn't a law or an algorithm. It's the electorate making it clear that crossing personal lines costs votes instead of winning them.
Campaigns will keep pushing the envelope as long as it works. The moment an ad like this backfires at the ballot box, the prompts will stop.
