Ahead of Knesset elections, Israel must enforce rules to combat AI disinformation - editorial

Israel's election campaign has offered a stark preview of the challenges ahead, particularly concerning the use of artificial intelligence to generate disinformation. A recent example involved an AI-generated Likud campaign video that depicted Gadi Eisenkot turning away from a figure wearing an "Israeli unity" shirt and embracing Ra'am leader Mansour Abbas.

Critics pointed out that the figure resembled Eisenkot’s son, Gal, who was killed in Gaza in December 2023. While Likud denied any intentional resemblance and later released a revised version, the incident was widely condemned as cruel and reckless. It highlighted a broader problem: the ease with which AI can create convincing, yet false, images, voices, and videos.

In response to this growing threat, the Central Elections Committee has taken a preliminary step by issuing temporary regulations. Committee chairman Justice Noam Sohlberg mandated that campaign materials created or substantially altered using digital means must be clearly labeled. These disclosures need to be prominent, with labels remaining visible throughout video footage and warnings at the beginning and end of audio material. Failure to comply may be considered a legal violation.

However, the article argues that these labeling requirements are merely a beginning and are insufficient to combat the rapid spread of AI-driven misinformation. Labels can inform voters, but they cannot prevent misleading content from proliferating across social media platforms like Telegram, TikTok, X/Twitter, Instagram, and WhatsApp before regulators or journalists can intervene. Furthermore, the issue of accountability remains complex when deceptive material originates from unofficial accounts, surrogates, or supposedly independent organizations.

The editorial emphasizes that political parties must be held responsible for content disseminated through official channels and coordinated efforts. Campaigns should not be able to benefit from fabricated videos and then evade responsibility by blaming supporters. The Central Elections Committee needs to establish a swift and transparent complaint process for synthetic media, as election misinformation travels at digital speed. Penalties must be significant enough to deter rule-breaking, rather than merely being a cost of doing business. Social media companies also have a crucial role, not in censoring lawful speech, but in responding swiftly when clear evidence of fabricated content emerges.