Israel's first AI election

What has actually changed?

In conversation with Kikar HaShabbat, Idan Ring, vice president for community and society at the Israel Internet Association, outlined the threats posed by artificial intelligence and its products, which are especially important to recognize during election season:

False and misleading information: False claims, data stripped of context, partial quotes, or old information presented as current.

AI-generated, edited or altered content: Images, videos and audio recordings depicting events that never happened, created or manipulated using artificial intelligence to deceive.

Impersonation and source forgery: Fake accounts, fabricated screenshots, or messages impersonating reliable or official sources such as media outlets, candidates, or state and security bodies.

Coordinated campaigns and foreign influence: Networks of inauthentic accounts and bots operating in coordination, run by malicious or foreign actors attempting to influence discourse, create a false impression of mass support, amplify polarizing messages, and sow confusion and distrust.

Viral spread of unverified information: Rumors and unsubstantiated claims can spread rapidly by word of mouth, on WhatsApp and social networks, often long before anyone has checked whether they are true. This is especially the case during periods of crisis, conflict or widespread public interest, and particularly during elections.

AI tools as a source of information for voters: AI tools can present incorrect, outdated or incomplete information, even when the answer is phrased confidently and convincingly. Voters should not rely on recommendations, information and content provided by chatbots like ChatGPT, Gemini and others.

The new risks

To understand the changes in public opinion during elections and beyond, we asked our sources: What risks have been added to running an election campaign since AI entered the picture?

Taking a more critical and forceful approach is Tomer Avital, co-founder of Shakuf and Lobby 99, a lecturer on critical thinking, author of the guide "Truth or Fake" and a candidate for the Democrats party. He told Kikar HaShabbat that "AI in elections has shattered discourse and reduced it to zero. And if that's not enough, we've reached a situation where no one knows what to believe, and even when something is real, it's easy to claim it's fake."

According to him, "Labeling content as 'AI' is a band-aid. The deep solution is education in critical thinking from a young age: teaching how to identify a reliable source, distinguish between fact and opinion, and understand framing and propaganda. Not telling people what to believe, but giving them tools to think for themselves."

So has artificial intelligence ruined everything?

Given these stark assessments, we asked our sources whether there are any positive developments that have come thanks to artificial intelligence, and whether hope for a genuine and normal election campaign has been lost.

Bar Zik noted that "artificial intelligence makes it easier for parties to analyze data and mobilize voters, but creates privacy dangers. AI election broadcasts have become more entertaining, and it's easier to expose politicians' lies."

Avital pointed to positive civic initiatives thanks to AI, such as games to encourage voting and platforms comparing party platforms. "But when you put that against the damage and the degradation of discourse, the minuses outweigh the pluses."

The Central Elections Committee's activity in the field

The Central Elections Committee told Kikar HaShabbat that the main challenge today is distinguishing between reliable information and misleading content, in an era when artificial intelligence makes it easy to create convincing content.

Therefore, the Committee operates a professional team dealing with illegitimate attempts to influence the elections, cooperates with platforms and works to educate the public. According to them, the message to voters is simple: when it comes to the election process, the address for authorized information is the Committee's official publications.

The Committee also notes that the change is also in the way we search for information: instead of searching a website, many ask an artificial intelligence model. But an answer that sounds convincing is not necessarily correct or current. Therefore, when it comes to the election process, it is important to check the answer against the Committee's official information.

At the same time, the Committee says that AI has also brought positive changes. For example, artificial intelligence can make complex information simpler and more accessible, in different languages and adapted to different needs. From the Committee's perspective, this has the potential to help more people understand the election process and exercise their right to vote. In conclusion, the Committee says that "the condition is that the information be accurate and reliable, so that the choice itself remains in the hands of the voter."

How to avoid illegitimate influence

Check the source: Who published the information? Look for the original source. A screenshot or WhatsApp message is not enough. Check whether it is an official body or an anonymous profile, and whether the publisher has an interest in influencing.

Prefer authorized sources: The Elections Committee, government websites, the Knesset, and the official accounts of parties and candidates.

Verify with additional sources: A significant claim? Don't settle for one post. Check whether the information appears in other reliable sources. Remember: ten posts are not ten sources, if they all quote one source, it's not enough. Distinguish between fact and personal opinion.

Check date and context: A short clip can change meaning. An old image can be presented as new. AI tools make it possible to fake images and videos, don't decide based only on "looks real." Look for the source and verification. Content created by AI is supposed to receive prominent labeling.

Proper use of artificial intelligence

If using artificial intelligence, only for initial research and finding sources, not for making decisions. Formulate clear and detailed requests, ask for official sources. But even a good prompt is no guarantee, open the sources the chat refers you to and check that they exist, are reliable and support the claims. Cross-reference with additional sources you find yourself.