Voting in elections is not like searching for a recipe, drafting an email, or getting help with homework
A new survey by the Israel Internet Association reveals that a quarter of the public is considering consulting an AI chatbot ahead of the elections. The article warns against false authority, outdated information, inherent biases, and deliberate data poisoning, emphasizing that the decision of whom to vote for must remain with the voter.
Ahead of the upcoming elections, a new survey by the Israel Internet Association reveals that a quarter of the Israeli public (25%) is positively considering or intends to use an AI chatbot such as ChatGPT or Gemini to help decide their vote. The article, by Dr. Shaked Davran-Zivan, the Association's literacy and online safety coordinator, warns of the dangers of relying on such tools in a political context. It explains that chatbot responses may be perceived as objective and factual, but in reality they are influenced by the data on which the system was trained and the way the question is phrased. The article cites a 2026 British study in which most users perceived political information from AI as accurate and neutral, and warns that this tendency can be misleading. It also mentions a case study from the Netherlands, where chatbots tended to recommend only two out of 15 parties. Alongside unintentional errors, there is a risk of deliberate manipulation and data poisoning by interested parties. The article calls on voters to first turn to independent sources such as party platforms and official websites, and to use AI only as a starting point while applying critical thinking.