Central Elections Committee to work on combating AI influence on vote

The Central Elections Committee said Monday that it was setting up a dedicated team in conjunction with the Shin Bet to examine concerns about AI influence on the upcoming national election.

Speaking during a hearing of the Knesset Science and Technology Committee, Orly Adas, the director general of the CEC, said the body works with many security agencies in Israel ahead of each election campaign to ensure a safe and secure election.

“We are in the process of locating consultants in the field of AI who will assemble a dedicated forum on the issue, and it is on the agenda,” Adas told the committee, according to the Knesset spokesman’s office.

Adas said that the upcoming election is the sixth national vote in which “we will convene a very broad forum that includes all security bodies in the State of Israel, including the Shin Bet, the National Cyber Directorate, the Privacy Protection Authority, and others, in order to make use of tools and insights that only they can provide.”

She said that the CEC works regularly with the Shin Bet, and has held a joint meeting and “established a dedicated team” to address such threats ahead of the upcoming vote.

Adas admitted that the committee faces an uphill battle in combating the spread of false information online: “It’s not feasible that we’ll be able to clean up the online space of problematic and fictitious campaigns and content — even if a large campaign is removed, it can reappear within hours.”

While there is no current date for the next national election, it must be held by October 27, 2026. Since the last election, in 2022, AI generative tools have exploded in popularity, flooding social media with fake images and videos.

Even Prime Minister Benjamin Netanyahu has shared such media, including in October, when he posted an AI-generated image on his official prime ministerial X account depicting US President Donald Trump receiving a Nobel Peace Prize.

Trump himself regularly posts and shares such videos, including one last February which showed himself and Netanyahu lounging poolside at a “Trump Gaza” hotel in the rebuilt Strip.

While many such videos are obvious fakes, as AI tools improve their capabilities, many videos and images are becoming indistinguishable from real ones.

Speaking at the Knesset committee hearing on Monday, Lior Weintraub, the director of public policy at TikTok Israel, said the platform had tools to clearly label and identify AI content.

“During election periods we open a war room in order to provide immediate responses to all issues that arise,” said Weintraub. “We developed an automated tool based on AI that creates a label indicating that the content was created using AI,” he added, claiming that it has been 97% effective in flagging or removing such content.

Idan Ring, a representative of the Israeli Internet Association, said platforms can and should be doing much more to counter misinformation.

“Fraud operations are happening right now under our noses. We are seeing [fake] videos of ministers recommending stocks” and other campaigns that “include Iranian involvement,” said Ring.

Ring accused social media platforms of being largely unresponsive to the IIA’s concerns: “We need to demand much more, we need to demand dedicated preparations,” he said. “The platforms have tools, they know how to stop campaigns that reappear. We need to demand more answers.”

A study published last month in the Science and Nature journal by a Cornell University professor found that a brief conversation with a partisan AI chatbot can influence voters’ political views, with evidence-backed arguments — true or not — proving particularly persuasive.

The studies found that the most common tactic used by chatbots to persuade was “being polite and providing evidence,” and that bots instructed not to use facts were far less persuasive. But the facts and evidence cited by the chatbots were not necessarily truthful, the study found.

AFP contributed to this report.