Quarter of Israelis may use AI to decide how to vote, survey finds

The association warned that disinformation, impersonation, and abusive uses of AI have become part of the current online environment and could intensify during the campaign, potentially reinforcing one another.

By Pesach Benson, TPS

A quarter of Israelis are considering using artificial intelligence to help decide how they will vote in the upcoming election, while widespread exposure to online disinformation is raising concerns about manipulation and political violence, a new survey found.

The survey, conducted by the Geocartography Group for the Israel Internet Association and released on Sunday, found that 25% said they were considering or intended to use AI or another online service to help make their voting decision.

Elections are scheduled for October 27.

At the same time, 71% said they had encountered false election-related information frequently or occasionally. Only about 15% said they had not encountered such information.

Fake or impersonating accounts were also widespread, with 56.8% reporting that they encountered them frequently or occasionally.

Nearly half, 48.2%, said they had encountered AI being used to harm or defame someone, while another 25.6% said they had encountered such use rarely.

Concern over the potential impact of online political discourse was similarly high.

About 69% said they were very or extremely concerned that the election campaign could lead to increased violence on social media and other digital platforms.

Exposure to election disinformation was reported across the political spectrum, with 73.3% of right-wing respondents saying they had encountered it frequently or occasionally, compared with 69.5% of centrists and 68.9% of left-wing respondents.

The association warned that disinformation, impersonation, and abusive uses of AI have become part of the current online environment and could intensify during the campaign, potentially reinforcing one another.

It urged voters to verify who is behind information, favor authoritative sources, compare claims across multiple outlets and check publication dates and context.

The association also cautioned against treating AI chatbots as objective authorities. Such systems can provide useful tools for comparing party positions and researching issues but may rely on outdated or incomplete information, confuse fact with interpretation, or omit context or fabricate facts, quotations, and sources.

The poll was based on 1,005 respondents representative of the Israeli population.