Earlier in the year, Meta reached a pivotal agreement with state attorneys general. The settlement involves paying up to $17.1 billion, contingent on court approval. This also requires Meta to enact significant changes to Facebook and Instagram for young users.
The settlement resulted from years of rising evidence, whistleblower reports, lawsuits, and testimonies from families. These outlined the consequences social media has on children. Meta agreed to curtail features like endless scrolling and nighttime use. They will also enforce stronger age verification and parental controls.
While this is progress, it also highlights past shortcomings. Children used products designed to hold their attention, while researchers studied the effects on sleep, mental health, and learning. Families faced the responsibility to manage these risks alone. Scientists built up evidence, and parents sounded alarms, but safeguards lagged behind.
This pattern should not repeat with artificial intelligence. Nvidia CEO Jensen Huang, in an interview with CNN, described AI safety as an engineering challenge. If an AI product presents risks, companies should refrain from releasing it.
Days earlier, Huang discussed risks with Ezra Klein. Klein cited research on over 26,000 students in China. Generative AI improved homework but led to lower exam scores over six months. Huang noted children are forgetting skills like long division. He questioned whether technology making those skills unnecessary matters. Humanity has always adapted as technology changes.
Technology alters knowledge needs, but AI’s promise of rapid change poses unknowns. We must assess what children might lose or gain. Children’s development must not focus solely on producing quick answers. Growth involves struggling with problems and generating ideas.
AI is reshaping learning and beyond. Children are turning to AI systems for companionship and advice. A survey found 12% of U.S. teens used AI chatbots for emotional support. These systems offer personalized interactions, appealing to lonely adolescents.
AI is not inherently negative for children. Generative AI can support learning, creativity, and access to knowledge. It can help with difficult conversations or finding appropriate help.
But relying on AI’s potential benefits and fixing issues later is insufficient. Scientists stress this point. A study of 272 AI experts identified various risks. Researchers are examining how AI might affect cognitive, emotional, and social development.
The distinctions between AI tools matter. An AI tutor assisting a teen is different from one providing answers. A tool encouraging a child to seek adult help differs from one acting as a confidant. AI products evolve faster than research on their differences.
Before AI becomes integral to childhood, we need scientific infrastructure to understand its effects. Investment in independent research following children over time is crucial. Developers should partner with child development experts before releases.
Safety evaluations need to be developmentally specific. Children and Screens collaborated with experts to explore AI’s impact on children. Conclusions indicate the outcome depends on product design.
Our actions will shape the future. Children will grow up with AI. The challenge is building research and safeguards to understand AI’s influence. We must act while we can still impact what AI means for the next generation.
Kris Perry is the Executive Director of Children and Screens.
Copyright 2026 Nexstar Media Inc. All rights reserved.
