Israelis Increasingly Turn to AI Chatbots for Voting Decisions
A survey reveals 25% of Israelis plan to use AI chatbots for voting decisions. Experts warn of algorithmic biases, data poisoning, and the necessity of critical digital literacy.
A new and unconventional player is set to enter the political arena ahead of the upcoming elections. While this entity will not appear on the ballot, run on any party list, or address crowds at rallies, it may well influence how a significant number of Israelis decide which ballot to cast: the artificial intelligence chatbot.
Most Israelis regularly consult chatbots for virtually everything—drafting emails, planning trips, studying, handling work matters, and even seeking relationship advice. As election day approaches, however, politics is inevitably entering the chat. Many Israelis intend to ask ChatGPT, Gemini, and other AI tools questions such as "Which party best aligns with my views?", "Which candidate actually advanced the issue that matters to me?", and ultimately, the most direct question of all: "Who should I vote for?"
The Allure of Algorithmic Objectivity
A new survey by the Israel Internet Association reveals that a quarter of the Israeli public—25%—is actively considering or intending to use these services to help make their voting decisions. However, turning to AI for political consultation is hardly equivalent to searching for a recipe, drafting an email, or asking for help with homework. While a chatbot's response may appear measured and objective, it is fundamentally shaped by the training data and the phrasing of the prompt.
Research shows that humans tend to attribute high objectivity and reliability to information delivered by machines, operating under the assumption that they are free from emotion or hidden interests. Yet this impression is deeply misleading. The fact that a response is presented in a neutral or authoritative tone by a machine does not mean it is accurate or free from bias.
This tendency to trust bots is reflected in a 2026 survey of adults in the UK, where the majority of users rated the political information they received from AI tools as mostly or entirely accurate, largely perceiving the answers as politically neutral.
Echo Chambers and the Risk of Misinformation
The conversational nature of AI further exacerbates the risk of deception. The more we interact with a chatbot, the more it tailors its output to our specific interests, thought patterns, and phrasing. When the conversation turns to politics, this personalization risks reinforcing pre-existing biases and assumptions rather than exposing users to challenging information or arguments that might prompt them to reconsider their positions.
Furthermore, AI systems carry inherent risks: they may rely on outdated data, confuse facts with interpretation, omit vital context, attribute outdated positions to political parties, and occasionally "hallucinate" entirely fabricated facts, quotes, and sources that sound completely convincing.
"During last year's parliamentary elections in the Netherlands, the Dutch Data Protection Authority warned against using chatbots for voting advice after finding that they tended to recommend only two out of 15 parties, presenting a partial and polarized picture of the political landscape."
Beyond unintended errors and hallucinations, a more deliberate threat looms: intentional manipulation and "data poisoning." As more voters turn to AI for political information, the incentive to tamper with the data underlying these models grows. Interested actors may attempt to poison models with false or misleading information to sway voter behavior—tactics already attributed to state actors like Russia, which has utilized fake research institutes and heavy online publishing to influence global public opinion.
Critical Literacy in the Age of AI
While public discourse surrounding elections and AI frequently focuses on deepfake videos, altered images, and widespread disinformation, the true challenge of this election cycle may lie in the quiet, private conversations we conduct with a system perceived as a reliable and decisive personal assistant.
This does not mean artificial intelligence has no place in the research and comprehension of the political map. It can help untangle complex issues, map out disputes, or compare political platforms. Yet when seeking information to form a political stance, citizens should first turn directly to primary sources: party manifestos, official websites, public data, academic studies, and professional journalism.
If voters nevertheless choose to utilize AI, they must cultivate new tools and skills in digital literacy and critical thinking. Users can ask the chatbot to present arguments for and against a specific policy or direct them to official sources detailing party positions. However, the resulting answer must serve merely as a starting point.
It is essential to inspect the cited sources, verify who stands behind them, check whether they are current, and ensure they genuinely support the claims made—subsequently cross-referencing the information with reliable independent sources.
The upcoming election will be the first in Israel—and among the first globally—where chatbots play a substantial role in political discourse and voter decision-making. While it remains too early to measure their exact impact on final results, with one in four Israelis contemplating asking a chatbot how to vote, these systems can no longer be treated as ordinary web sources. AI can help navigate data, but it is no substitute for primary sources—and certainly no substitute for individual human judgment. On the road to the ballot box, citizens must read, verify, and cross-reference information independently.
