As artificial intelligence rapidly permeates daily life, students and adolescents face an unprecedented cognitive environment: information can be generated at near-zero cost, algorithms deliver content with high precision, and social and companionship functions are highly commercialized. This makes critical thinking and critical learning more crucial than ever, and explains why in the past year or two many governments and legislatures have begun seriously considering or even implementing limits on social media and certain AI tools for minors under 15/16.

1. AI lowers the barrier to "getting answers" while raising the barrier to "judging answers"

Generative AI can produce writing, solutions, summaries, analyses and even emotional dialogue in seconds. For students, this is both a powerful learning assistant and a potential replacement for thinking. Without critical skills, there is a risk of a false sense of understanding—accepting biased or hallucinatory outputs, skipping derivations, or taking algorithmically recommended views as one's own judgments.

Critical thinking requires the habitual questioning of: where does the evidence come from? Are there overlooked counterexamples? Does this conclusion merely reflect how I asked the question? Critical learning goes further, emphasizing understanding how knowledge is constructed, whether independent verification by other methods is possible, and how to treat AI as a tool rather than an authority.

For adolescents whose executive functions and metacognition are developing rapidly, chronically outsourcing thinking to AI may weaken abilities to analyze independently, spot problems and build arguments. If these capacities atrophy during critical developmental windows, remediation later will be far more costly.

2. Real risks from social media and AI are driving policy shifts

In the past year or two, Australia took the lead in imposing nationwide social media account restrictions for those under 16; Brazil, Indonesia and Malaysia followed, and the UK, France, Greece, the UAE, Denmark and some US states are advancing similar measures, while also bringing AI chatbots and companion tools into minors' protection frameworks. Major drivers include:

Mental health pressure: infinite scrolling, algorithmic recommendations and social comparison are linked to anxiety, depression, body-image issues and sleep disorders;
Harmful content and safety risks: cyberbullying, inappropriate material, self-harm related content and risks of predatory contact;
Addictive design: features intentionally extending time spent on platforms, crowding out real-world socializing, study and outdoor activity;
AI-specific problems: emotional dependence, inappropriate advice and large-scale collection of minors' conversational data.

These policies essentially acknowledge a structural conflict between highly optimized commercial algorithms and still-maturing adolescent brains. Exposing youth to intense digital stimulation before protective systems (judgment, impulse control, long-term planning) are in place significantly raises risk.

3. Limits are the "guardrails," critical skills are the "steering wheel"

Purely prohibitive measures cannot solve the problem. No ban is fully foolproof; more importantly, adolescents will eventually access more open digital environments. If we only impose limits without cultivating judgment, opening access later may lead to loss of control; if we only promote openness without safeguards, development during key periods may suffer irreversible harm.

Both approaches must be combined:

Early emphasis on protection: restrict highly addictive features, enforce effective age verification, default to safe settings, and set stricter thresholds for AI companion tools;
Simultaneous, systematic cultivation of critical skills: train students at school and at home to understand algorithmic logic, verify AI outp