As Israel prepares for future electoral cycles, government efforts to regulate artificial intelligence are currently centered on the detection and mitigation of deepfakes—manipulated videos or audio clips designed to impersonate political figures. Officials argue that these visual deceptions pose the most immediate threat to democratic processes by creating false evidence of candidate behavior.

However, cybersecurity experts and political analysts suggest that this focus is insufficient. They contend that AI-powered chatbots, such as ChatGPT and Gemini, present a more pervasive and subtle danger. Unlike deepfakes, which are often easily debunked once identified, chatbots can engage in personalized, large-scale interactions with voters. These tools have the capacity to sway public opinion, suppress voter turnout through targeted messaging, and generate vast amounts of content that challenge the legitimacy of election results.

The core disagreement lies in the scope of regulatory oversight. While the state is prioritizing visual misinformation, critics argue that the ability of language models to automate propaganda and disinformation campaigns is being overlooked. The concern is that by focusing exclusively on the 'video' aspect of AI, the government is leaving the electorate exposed to more sophisticated, text-based manipulation that is harder to track and regulate.