By Ran D. Anbar, MD, and Ayush Prakash
Note: This article was reprinted with permission of the author.
You can read the original article here.
Personal Perspective: The Case for Limiting Exposure to AI and Social Media
In previous blogs, I have discussed how AI and digital technology might be used to enhance education, but can also be harmful. In this blog we will explore how overuse of AI in education and exposure to social media may affect the brain, which typically continues to develop until around the age of 25 years.
To best understand how AI affects brain development and therefore education, we must first review the nature of AI systems in current use. The most prominent AI systems today are Large Language Models (LLMs) like ChatGPT, Claude, Grok, Perplexity, and Gemini.
These systems work through computational models that mimic the human brain’s structure, thus termed “neural networks.” They consist of interconnected nodes that process and learn from internet data, enabling pattern recognition and decision-making in the field of artificial intelligence called “Machine Learning.” LLMs are trained on massive datasets containing billions of words from books, websites, and other text sources.
Critically, LLMs do not “understand” anything in the ways humans do. Simply, they identify statistical patterns. When an LLM generates a response, it is essentially completing a pattern based on similar contexts it encountered during training (Bender, et al., 2021).
This means that in addition to providing information, LLMs can reproduce biases present in their training data, generate plausible sounding but incorrect information (a phenomenon called “hallucinations”), and completely lack the ability to synthesize genuinely novel ideas that go beyond recombination of existing patterns. (Chollet, 2019)
Perhaps more consequential are the recommendation algorithms that power social media platforms, video streaming services, and content feeds. These systems are similarly a product of the field Machine Learning and predict what content will keep users engaged longest. The algorithms analyze vast amounts of data about user behavior (what they click on, how long they watch, what they share, when they pause or scroll) to build predictive models of individual preferences.
The optimization target for these algorithms is typically “engagement,” measured through metrics like time on site, clicks, shares, and comments. This creates a fundamental misalignment. The social media algorithm’s goal is not to inform, educate, or promote wellbeing, but rather to maximize attention capture.
Such is becoming the case for products like ChatGPT, Claude, Perplexity, Grok, Gemini, etc. who are beginning to embed ads in their software (Search Engine Land, 2023) which simultaneously enhances their Time-On-Site metrics for users (a large number of which are children).
Harvesting Youth
To understand AI and social media’s potential impact on brain development, we must recognize that most of these systems that children encounter are embedded within an “attention economy” where human attention is the scarce resource being competed for and monetized.
The psychological mechanisms these systems employ to capture attention are not emergent or accidental, rather they are the product of deliberate design informed by decades of behavioral psychology, data science, and neuroscience research. Social media platforms, for example, employ several mechanisms specifically designed to create impulsive and compulsive use patterns.
Like slot machines in casinos, social media feeds provide unpredictable rewards – a funny video, an interesting post, a notification from a friend. The uncertainty of what might appear next keeps users scrolling. This unpredictability is more psychologically powerful than consistent rewards because it prevents the natural satiation that occurs with predictable outcomes. (Montag, et al., 2019) Every scroll might reveal something interesting, so users keep scrolling.
The aforementioned strategy works by hijacking the brain’s reward system. Each “like”, notification, or piece of compelling content triggers a release of dopamine, the neurotransmitter associated with pleasure, motivation, and learning. The brain learns to associate checking the device or scrolling the feed with potential reward, creating habit loops that operate below conscious awareness.
Platforms that display likes, followers, shares, and view counts tap into fundamental human needs for social acceptance and status. For adolescents, whose identity development is particularly tied to peer evaluation, these metrics become disproportionately important. The anxiety of “How many likes did I get?” or “Why didn’t they respond?” keeps young people checking obsessively.
By analyzing billions of data points about user behavior, algorithms create highly personalized content feeds that show users more of what they’ve previously engaged with. This reduces cognitive dissonance and increases engagement but creates echo chambers where users are rarely exposed to perspectives that challenge their existing views. For children developing their understanding of the world, this can create a false sense that their particular perspective represents universal truth.
On a relevant tangent, adolescents comparing themselves to AI-filtered and AI-enhanced images on social media face a reality where the beauty standards they aspire to are literally impossible for actual human bodies to achieve. The psychological harms of social comparison to curated and edited images are compounded when the images represent synthetic rather than even highly edited human appearance.
As AI-generated content becomes more prevalent and sophisticated, distinguishing synthetic from authentic media becomes both more important and more difficult. Children must develop a healthy skepticism about media they encounter while not descending into a situation where no evidence is trusted. Navigating this balance requires metacognitive sophistication that develops only gradually.
Vulnerability In Developing Brains
The prefrontal cortex, responsible for planning, impulse control, and self-regulation, undergoes substantial development throughout childhood and adolescence, typically not reaching maturity until the mid-20s.
This means children and teenagers have structurally less capacity for self-regulation than adults, and are less able to resist compelling stimuli, less able to recognize when their technology use is excessive, and less able to disengage from rewarding activities even when they intellectually understand they should. AI and social media systems optimized to be maximally engaging are fully exploiting this developmental vulnerability.
During adolescence in particular, the brain’s reward system undergoes changes that create heightened sensitivity to rewards, especially social rewards. This is an adaptive feature that motivates adolescents to form peer relationships and take risks necessary for independence. Also, importantly, it makes teenagers disproportionately susceptible to the social validation mechanics of social media and the dopamine loops of engagement-optimized algorithms.
The neurochemistry that should drive healthy risk-taking and social exploration is hijacked by artificial reward systems that trigger similar neurological responses without serving developmental needs.
Critical windows exist when the brain is maximally plastic to acquire certain capacities. The neural pathways that are strengthened through repeated activation during these periods become permanent brain architecture. Pathways that go unused are pruned away during later childhood.
When children spend substantial portions of critical developmental periods engaged with AI systems and digital technology, they are strengthening neural pathways for rapid content switching, shallow information processing, and superficial technosocial interactions while failing to adequately develop pathways for sustained attention, deep reading, face-to-face social interaction, and effortful problem-solving in the real world. After the critical period closes, these underdeveloped capacities may be permanently diminished (Tooley et al., 2021).
Conclusion
Based on available evidence, we believe that AI and social media use with children and adolescents should be minimized, if not eliminated, until we develop a better understanding of their impact on the developing brain.
Ayush Prakash is the Creative Director of New Sapience, author, and podcaster focused on the cultural and societal impact of AI. He is the author of AI for Gen Z and host of the Ayush Prakash Podcast, a platform for critical, intergenerational dialogue on technology, identity, and the future.
