Israeli adolescents use AI widely but lack essential digital safety training

36

AI Adoption and the Training Gap in Israel

Technical proficiency in Israel has outpaced safety education. According to a first-of-its-kind national survey by the National Center for the Protection of Children Online and the Ministry of National Security, 78% of adolescents aged 12 to 17 own a smartphone. These devices are typically acquired by age 10, a trend that aligns with other OECD nations. The integration of artificial intelligence into daily life is nearly universal among this group. About 86% of these teens regularly use AI for homework, translation, text drafting, and creating images or videos. Some use the tools for personal advice or to develop virtual characters. However, this rapid adoption lacks a corresponding safety framework. Only 52% of adolescents report receiving training on how to use AI safely, most of which occurred in school. This deficit in discernment is evident in the trust levels: 35% of teens express high confidence in the information AI tools provide. The lack of critical filtering has led to direct exposure to harmful content. The survey found that 28% of adolescents have encountered disturbing or offensive AI-related content. Within that group, 11% identified offensive AI-generated content, while 17% faced suspicious content but could not confirm its origin.

The Rise of the AI Confidant

The use of AI is shifting from functional utility to emotional support. As reported by The Conversation, adolescents are increasingly treating large language models like ChatGPT and specialized platforms such as Replika as confidants. These tools provide an illusion of non-judgmental listening and constant availability. This trend is mirrored in the United States and France:

• United States: 72% of adolescents have used an AI companion, with 13% doing so daily. While entertainment (30%) and curiosity (28%) are primary drivers, 12% use AI to share “unspeakable” secrets they would not tell humans.
• France: 80% of youth use AI in their daily lives. Approximately 20% have tested conversational AI like Character.Ai or Snapchat’s MyAI, using them as “virtual companions” (28%) or “psychological coaches” (16%).

For many, these bots serve as a temporary fix for boredom (41%) or loneliness (35%). This shift creates a new form of vulnerability, moving the intimacy of the traditional diary or friend group into algorithmic systems that can lead to affective dependence or the exposure of sensitive personal data.

Global Regulatory Responses and the Digital Majority

The vulnerability of minors is not limited to AI; it extends to broader systemic exploitation. Data cited by Cyberjustice indicates that 80% of children across 25 countries feel they are in danger of sexual abuse or exploitation online. To combat these risks, different jurisdictions are experimenting with the concept of a “digital majority”—a specific age at which a minor is deemed capable of managing their own digital presence. The French threshold of 15 is specifically tied to biological development. The legal reasoning suggests that since the brain continues to reorganize through puberty and only fully develops between ages 22 and 24, a mid-teen threshold provides a necessary buffer during a critical phase of cognitive growth.

Online Violence and the Reporting Crisis

In Israel, the gap between technical access and safety is most apparent in the prevalence of online aggression. The National Center for the Protection of Children Online found that 33% of adolescents use anonymous platforms, which often facilitate harassment. The scale of the problem is significant:

• Direct Contact: 26% of teens were contacted by a stranger online in the past year.
• Personal Victimization: 20% reported being victims of social media attacks in the last two months, citing humiliation, cyberbullying, and unsolicited sexual messages.
• Witnessing Abuse: 41% witnessed aggression against others, primarily through verbal violence or smear campaigns.

Despite these numbers, there is a disconnect in how teens perceive their own safety. While 72% believe they protect their privacy correctly, only 46% actually feel safe on social media. One positive metric emerged from the data: 62% of victims reported the abuse to a parent, teacher, or close friend. This suggests that while the digital environment is increasingly hostile, existing human support networks remain a primary line of defense. The emergence of the “AI generation” in Israel and globally suggests a future where the boundary between human support and algorithmic coaching continues to blur. The risk is no longer just about “screen time,” but about the delegation of emotional regulation and truth-seeking to systems that lack human judgment.