How AI is becoming a new battleground in Israel's election

East Jerusalem - Like a scene from an action movie, Israeli political party leaders Gadi Eisenkot and Avigdor Lieberman descended an elevator shaft, tore off their spy gear, and tried to break into the prime minister’s office.   

When their attempts failed, Eisenkot said to Lieberman in Hebrew, “I have the master [key],” and the video cut to Israeli MP and leader of the United Arab List (Ra'am), Mansour Abbas, waving a giant set of pliers with a sinister grin.

In fine print in the top right-hand corner of the video, it reads 'AI: This content was created through a digital medium and is not original.' 

Israeli Prime Minister Benjamin Netanyahu posted the AI-generated video on his X account with the caption, “Eisenkot will form a left-wing government with [Israeli Democrats’ leader] Yair Golan and the Arab parties”.  

It’s one of a myriad of images Netanyahu and other Israeli politicians have shared ahead of Israel’s national elections scheduled for 27 October.

In fact, Netanyahu’s Likud Party leads in the amount of AI-generated content they’ve distributed throughout this election cycle.

According to the “Ze Be AI” project by Israel’s Open University, 13 per cent of content published by politicians and political parties was created with artificial intelligence. Of that number, 70 per cent of the AI content came from Likud.  

Highlighting their opponents’ willingness to work with Palestinian parties is a common theme for these AI campaign adverts.   

AI becomes a new political weapon in Israel

In a video posted by Itamar Ben Gvir, Israel’s National Security Minister and leader of the Jewish Power Party, former police officer Yoav Segalovitz is seen wearing the flag of the Al-Qassam Brigades, Hamas’ military wing, while handing out pastries to a smiling queue of Palestinian prisoners.

In the preceding scene, Abbas kisses Segalovitz while draping him in the green sash. Segalovitz joined Abbas’ party last month, prompting Ben Gvir to release the AI-generated video with the caption, “Marhaba (hello in Arabic), ya Segalovitz. The Eisenkot brothers and [former Israeli Prime Minister Naftali] Bennett are proud of you,” in Hebrew.  

Another AI video, posted by Bezalel Smotrich, Israeli Finance Minister and leader of the Religious Zionism party, portrays a phone conversation between a Palestinian in the occupied West Bank and a Palestinian citizen of Israel, with the West Bank resident warning that Smotrich is going to Judaize the Naqab (Negev) and Galilee, just as he did with the West Bank. 

“Over here it is all lost. They don’t let you run wild anymore and build without stopping,” the character, Ahmed, says. 

Using AI, Smotrich, Ben Gvir, and Netanyahu are attempting to instil fear in Israeli voters that their opposition - in collaborating with Palestinian parties - will lead to terrorism, and that under their leadership, Israel’s expansionism will thrive.  

In response to several controversies surrounding the use of AI-generated campaign materials, including a Likud video showing Eisenkot running into the arms of Abbas after snubbing an individual resembling his son, who was killed in Israel’s genocidal war on Gaza, authorities have stepped in.

Israel’s Central Elections Committee has mandated the aforementioned AI label to be displayed on all AI-created campaign content that “could appear to have been authentically documented when it was not,” meaning not all AI images have to be disclosed. 

“It's not real, it's an AI. So when we cannot tell what is true and what is generated any longer because of AI slop, people’s minds become very weary. They don't have the energy to start determining what is real and what is not,” Dr Sharon Haleva Amir, a communications professor at Bar-Ilan University, told The New Arab.

“What it causes is that you can no longer say what is true and what is false in the general sense -  to an ad or a statement or an agenda. You just don't believe anything anymore.” 

Changing the way Israelis vote

A month after Israel’s last election, in November 2022, ChatGPT debuted, and subsequently unleashed a swarm of AI tools into the public sphere.

Almost four years later, this is Israel’s first election in the age of AI and will test how much influence the technology has - not just over day-to-day lives, but politics as well. 

“These things meddle with your brain in ways that the people don't understand or perceive in real time,” Haleva Amir said. “From a democratic point of view, AI pulls the rug out from under your feet because this is not really your decision. Your mind is being played with.”  

And with AI distorting reality, Dr Tal Elovits, a political analyst and communications lecturer at Reichman University, explained it reinforces the scepticism that Israeli society was already experiencing in the aftermath of the coronavirus pandemic and Hamas’ 7 October 2023 attack. 

“Conspiracy theories that are blending with political campaigns and with the generated AI content, which looks good, have a tremendous effect,” Elovits said. “It makes the whole information environment become untrustworthy. For the general public, if you don't know where the truth lies, if you don't know how you can search for the truth, it makes political convincing and political competition very, very difficult.” 

In addition to increasing the level of distrust in the public, AI also risks compromising election integrity.   

“One of the affordances of AI-generated content is that you can prepare a lot of versions. Different people can receive different ads,” Haleva Amir said.

“Like we have with search engine optimisation, today there are also firms that do large language model [LLM] optimisation, and in this way, we cannot even know what people are being fed with because these things are behind the scenes.” 

AI has become so advanced that its messages are no longer simply tailored to individual commercial interests.

“We are no longer talking about a single ‘campaign ad’ for everyone, but an emotional pressure point that speaks directly to the fears, anger, or hopes of a specific individual,” Dr Tehilla Shwartz Altshuler, head of the Israel Democracy Institute’s in the Digital Age Program, wrote. 

 As politicians use AI to influence, Israeli voters are expected to use it in their decision-making.  

Through a survey, the Israeli Internet Association found that a quarter of Israelis are considering or intending to use AI to determine who they should vote for. 

“If the information is not correct and since people do not check anymore, then the [LLMs] can circumvent the way voters think,” Haleva Amir told The New Arab.  

While AI has sparked alarm in the run-up to the election, it also poses a threat on Election Day as robocalls could disseminate false information to voters.

Similar to Netanyahu’s infamous 2015 rallying cry that falsely claimed, “Arab voters are coming out in droves to the polls,” AI bots could spread propaganda designed to manipulate votes.

Possible violations, like these, of the Central Elections Committee’s new AI law can be reported to the committee through a petition and if the material does breach the law, the content will be removed. Experts stress, though, it’s a little too late once the material is published.  

“The content itself is already living online. It already has multiplied itself,” Elovits said. “So the damage was already done even before the committee was able to discuss if it's illegal.”

Amid weak election oversight, Haleva Amir advises that the public should become more informed on the capabilities and dangers of AI.

“Laws will not be enough or sufficient because technology always exceeds the law,” Haleva Amir said.

“The real answer to all these things should come from the people…by raising awareness and educating people and giving them tools to criticise and reflect and understand AI literacy.” 

Jessica Buxbaum is a Jerusalem-based journalist covering Palestine and Israel. Her work has been featured in Middle East Eye, The National, and Gulf News

Follow her on Twitter: @jess_buxbaum

Edited by Charlie Hoyle