AI and Privacy Concerns Dominate Israeli Election Campaigns

Israel's current election cycle is the first to significantly incorporate artificial intelligence (AI) into political campaign strategies. Beyond the well-known risks of deepfakes and misinformation, complex questions are emerging regarding the extent of voter data political parties possess, their right to build "political profiles" on citizens, and the use of AI to tailor messages to individual voters. AI's ability to rapidly generate realistic videos, images, and audio recordings is transforming campaigning, enabling parties to process vast amounts of voter data to infer preferences and craft personalized messages.

The current election system stands at the intersection of election and propaganda laws, privacy and data security regulations, and rapid AI advancements. The law must now address not only what can be published but also what information can be used to determine message targeting and creation. Israeli legislation has recognized the risk of "deep fakes", digitally altered content that appears authentic, by mandating clear disclosures for such material in election propaganda. The goal is transparency, allowing voters to distinguish between genuine and synthetic content.

However, labeling content as AI-generated does not grant a license for any material. While it addresses the transparency requirement for synthetic media, other laws, including privacy regulations, still apply. Using an individual's likeness, voice, or personal information can raise privacy issues, necessitating scrutiny of content origin, data sources, and context.

Political parties are permitted to access voter registry information for campaign purposes. Yet, this access does not grant unlimited usage rights. Direct marketing based on personal characteristics is subject to privacy laws, requiring disclosure of the sender, data source, and the option to opt-out. The use of unclear or illegally obtained data by parties raises significant questions.

AI's ability to infer information about individuals, such as political leanings or voting behavior, intensifies privacy concerns. The origin and purpose of data collection, whether directly provided or algorithmically inferred, are crucial. The sensitive nature of political profiling demands careful examination of data usage.

While tailoring messages to audiences is not new, AI enables unprecedented scale, precision, and automation. Legal challenges lie in the data sources used for segmentation, the legality of data usage, transparency, and whether data is used beyond its original purpose.

Parties are also obligated to ensure data security. Large databases, especially those containing information on a broad public, require serious attention to security risks and legal obligations. Breaches could lead to financial penalties or criminal liability. The increased number of involved parties during elections, including vendors and activists, heightens the need for access control and oversight.

Foreign interference using AI is a significant concern, as AI dramatically lowers the cost and time for creating large volumes of deceptive content. Hostile actors could generate and distribute tailored disinformation rapidly, potentially targeting specific demographics. The speed of content creation and distribution, often originating outside Israel, poses a challenge for regulation and timely legal intervention.

To navigate this landscape, voters are advised to maintain healthy skepticism towards unusual content, especially near elections. Checking content origin, disclosure labels, and corroboration from reliable sources is recommended. Voters should also inquire about the source of personal information used in political outreach and exercise their rights regarding data privacy. Ultimately, in the age of AI, personal data protection is intrinsically linked to a voter's ability to make independent decisions.