Holding Big Tech to Account - A Generation too Late
The historic US verdicts holding Meta and Alphabet liable for engineered childhood addiction show that public opinion and legal frameworks have permanently shifted. Lawmakers are finally holding big tech to account, but this intervention arrives a generation too late. The children who grew up as the test subjects for infinite scroll and autoplay algorithms are already adults dealing with the consequences.
Google have recently settled a childhood addiction case in the states with a 15 year old in Florida. This follows a jury in California earlier this finding that Meta and YouTube were liable for their platform’s mental health effect on users. A jury in New Mexico has also ordered Meta to pay $375m for misleading users over the safety of it’s platform for children. All of this feels right to me, but it’s a little too late.
We cannot afford to make the same mistake with artificial intelligence. While social media manipulated social validation, AI interacts with a child's cognitive development. These emerging risks are happening during critical developmental windows.
Current research and cognitive theories point to several specific ways AI could negatively affect the next generation of children:
The Erosion of Cognitive Friction: Generative AI tools give immediate, perfect answers to homework, creative blocks, or logic problems. This removes cognitive friction - the struggle required to learn how to research, cross-reference, and think critically. Children risk losing the capability to problem-solve independently when an AI assistant always bypasses the effort.
Parasocial Synthetic Relationships: AI companions are marketed as always-available, perfectly compliant friends. Unlike human peers, an AI friend never disagrees, never throws a tantrum, and never requires compromise. Relying on synthetic relationships risks stunting emotional intelligence, leaving children unprepared for the messy, unscripted friction of real human socialization.
The Authority Bias and Deepfake Reality: Children naturally trust authoritative voices. Because AI conversational tools speak with absolute confidence, younger users struggle to separate hallucinated misinformation from factual truth. This worsens as multimodal deepfakes become indistinguishable from reality, making it incredibly difficult for a child to build an accurate baseline of truth about the world.
Algorithmic Confirmation Bubbles: If you think social media recommendation algorithms were isolating, generative AI is could easily be worse. Personalised AI models can generate custom stories, responses, and educational content tailored entirely to a child’s existing biases or anxieties. This risks creating a totally insular reality, preventing exposure to diverse viewpoints or challenging ideas.
Lawmakers were reactive with social media, treating the psychological impacts as a theoretical problem until the data became undeniable. With AI, regulation must be proactive. We need strict guardrails on cognitive engineering before the next generation becomes another unintended experiment.
