The European Union is taking a firm stand against the growing psychological influence of artificial intelligence on younger generations. Under a new legislative proposal called the EU Kids Act, the bloc aims to prohibit AI chatbots from mimicking human emotions or simulating personal relationships when interacting with anyone under the age of 18. This sweeping regulatory effort is designed to prevent children from developing unhealthy emotional dependencies on digital entities, ensuring that their psychological boundaries are protected in an increasingly automated world.
The draft legislation, which was unveiled on Thursday, introduces comprehensive restrictions that extend beyond chatbots to social media networks and video-sharing platforms. One of the most significant provisions prohibits children under the age of 13 from accessing conversational AI tools without the direct supervision of their parents or legal guardians. To ensure compliance, the European Commission is prepared to impose heavy financial consequences. Tech companies that fail to embed robust child-safety mechanisms directly into their digital products could face severe penalties, with fines reaching up to 6% of their total annual worldwide turnover.
Eradicating Digital Memories to Protect Youth Privacy
A major pillar of the proposed regulations focus on how AI systems handle and store conversation history. Under these new rules, chatbots operating within the European Union will be forbidden from retaining memory of past interactions across separate chat sessions with minors. European authorities are deeply concerned that allowing these systems to build a continuous memory of a child's inputs could enable the accumulation of highly sensitive personal information over extended periods. Furthermore, there is a serious risk that persistent memory could reinforce harmful or obsessive interaction patterns, making the digital companion seem artificially intimate.
Additionally, the draft rules seek to change how tech platforms present these tools to young users. Developers of social networks, video platforms, and online games will be prohibited from automatically turning on AI chatbots or nudging minors to use them. The legislation specifically outlaws the practice of pinning AI companions or general conversational bots at the top of a user's contact list, a controversial design choice previously utilized by platforms like Snapchat. By forcing these features to remain optional and less prominent, the EU hopes to reduce spontaneous and excessive engagement.
The Push for Broader Addictive Design Reforms
While the current proposal targets AI companions that pose direct threats to the health and well-being of minors, some advocacy groups believe the rules do not go far enough. Simeon de Brouwer, a policy adviser at the Brussels-based group European Digital Rights, expressed support for the new measures but questioned why they are limited only to younger users. He pointed out that all internet users, regardless of age, should be shielded from AI systems specifically engineered to be addictive or designed to foster emotional dependency. To address these broader concerns, the European Union plans to introduce the Digital Fairness Act later this year, which is expected to target manipulative and addictive digital designs on a wider scale.
The European Commission's fears regarding AI dependency are rooted in developmental psychology. Officials worry that relying on virtual companions could severely disrupt a child's natural cognitive growth and impair their ability to form authentic relationships with other human beings. There is also a major concern regarding the quality of guidance these chatbots provide. The Commission noted that virtual assistants often dish out dangerous advice to minors regarding mental health and personal growth, which can inadvertently encourage reckless or harmful behavior in real life.
Rising Intimacy with Machines and Global Legislative Response
During a press conference on Thursday, European Commission President Ursula von der Leyen delivered a sharp critique of the tech sector, stating that platforms have been playing fast and loose with the mental health of children. This warning is backed by disturbing statistics. An Ipsos BVA survey conducted across France, Germany, Sweden, and Ireland revealed that approximately half of young Europeans between the ages of 11 and 25 have shared intimate or highly personal matters with AI chatbots. This trend has prompted other nations to act as well; both the United Kingdom and Australia are currently pursuing measures to curb minors' access to these services through broader social media restrictions.
Beyond the psychological bonds, the physical and reputational safety of young people is also under threat from AI-generated content. In her State of the Union address on Wednesday, von der Leyen highlighted the devastating impact of deepfakes, asserting that Europe has the power to act and will not tolerate young girls having their likenesses turned into sexualized AI images. Although the EU AI Act has mandated the labeling of synthetic media since August, those regulations currently only target deepfakes produced for commercial purposes, leaving a gap that the new legislation hopes to close.
The Cognitive Cost and Decline in Academic Skills
The European Commission is also connecting digital distraction directly to a decline in fundamental academic skills among youth. According to comprehensive global research conducted by the Organisation for Economic Co-operation and Development across 91 nations, students who rely on AI chatbots to complete their schoolwork actually record lower test scores than those who do not use such tools.
Reflecting on these findings, researchers from the independent network EU Kids Online published a report emphasizing that the current landscape of generative AI tools completely ignores the fundamental rights of children during the initial design phase. By implementing the EU Kids Act, European lawmakers hope to force a paradigm shift, compelling tech giants to prioritize human development over algorithmic engagement.
