J F Vernon Blog

AI rots your mind and constrains creativity

JV: Learning experiences create new neural pathways in the human brain. Significant engagement with AI will affect a plastic human brain. What studies have been undertaken or are underway to understand this? 

KAI: You’re right that learning—and by extension, engaging deeply with AI—can shape our brains through neuroplasticity. Researchers are indeed beginning to explore how sustained interaction with AI affects neural pathways, cognition, and behaviour. Here’s a look at what’s been studied so far:

1. MIT Study: AI-Assisted Writing and Brain Engagement

One of the most concrete investigations to date comes from MIT’s Media Lab. In a recent (preprint) study, 54 participants aged 18–39 were split into three groups: one writing SAT-style essays unaided, one using Google Search, and one using ChatGPT.

• Findings: ChatGPT users showed lowest neural engagement (measured via EEG), reduced executive control, and less creativity. They also produced more formulaic, similar essays and increasingly relied on copy-pasting over time.
  
• In contrast, the “brain-only” group demonstrated higher brain connectivity, better memory retention, and greater satisfaction with their work—even when later allowed to use ChatGPT.
  
• Inspired by these findings, the researchers are now expanding the scope—evaluating similar impacts in programming tasks, where early results appear even more worrying. 

2. Analyses on Cognitive Atrophy and “Extended Mind” Risks

Beyond the lab, scholars have reflected on broader implications of heavy AI use:

• A recent article on generative AI warns that relying too much on AI tools could lead to cognitive atrophy, including reduced memory, concentration, and analytical ability. The study even references the “Google effect,” where knowing information is readily accessible online weakens our memory retention.
  
• Similar concerns are echoed in coverage by The Guardian, which highlights the risk of diminished critical thinking and creativity, especially in younger users who offload too much cognitive effort to AI. 

3. AI Companions and Social Cognition: Human Behaviour and Emotional Bonds

Some studies delve into how AI influences our social and emotional life:

• A report on teen use of AI companions—like Replika or Character.ai—reveals that while such apps can offer constant, judgment-free interaction, heavy usage raises concerns about impaired social skills, loneliness, and unrealistic expectations for human relationships. Long-term effects remain unknown.
  
• The concept of artificial intimacy examines how people can develop emotional bonds with AI—albeit in one-sided, anthropomorphic ways. Emotional reliance on AI may increase social isolation or dependency on these artificial interactions. 

4. Brain–Computer Interfaces (BCIs) and Neuroplasticity

In parallel with AI as a conversational tool, there’s intensive research into AI-linked technologies that directly interact with neural activity:

• Emerging BCIs, like those by Stanford (inner speech decoding) and Precision Neuroscience (Layer 7 cortical interface), are using AI to interpret or translate brain signals—sometimes aiming to restore communication or motor functions. These efforts have strong implications for neurorehabilitation and demonstrate plastic brain responses to AI-mediated interfaces.
  
• On the academic frontier, MindSpeech employs non-invasive fNIRS data and language models (e.g., Llama2) to decode imagined speech—another avenue where AI and brain signals intersect.
  
• Conceptually, frameworks like “brain co‑processors” envision AI working alongside the biological brain to reinforce plasticity, assist in recovery, or even augment memory and learning.
  
• Scientific progress continues on multiple fronts—some mapping how multi-sensory AI models mirror brain function to inform both neuroscientific experiments and AI design.

5. Neuroplasticity at the Adult Level—and AI’s Role

Human neuroplasticity doesn’t vanish in adulthood:

• A major breakthrough from the Karolinska Institute found neural progenitor cells in adult hippocampi, implying adult brains can still generate new neurons—an area key to learning and memory. AI-assisted methods were pivotal in identifying these cells.
  
• This discovery opens doors to therapies that encourage neuroplasticity through AI-mediated interventions.

Key Insights

ChatGPT & Cognitive Engagement

Lower neural activity, reduced memory/creativity; expanding into programming tasks.

Cognitive Atrophy & Extended Mind

Over-reliance on AI linked to weakening cognitive faculties; echoing “Google effect.”

Emotional/Behavioural Impacts

AI companions can help loneliness but may hinder social development or create dependency.

Brain–Computer Interfaces (BCIs)

AI decoding/restoring neural functions demonstrates brain adaptability and plasticity in rehabilitation.

Adult Neurogenesis & Plasticity

AI techniques reveal ongoing neuron formation—potential for neuroprotective or regenerative strategies.

What’s Next?

• Longitudinal Research Needed: Many studies so far are short-term or exploratory. It’s crucial to study long-term cognitive effects and brain changes from habitual AI use.
  
• Diverse Tasks and Contexts: We need data on how AI affects not just essay writing but programming, problem-solving, learning, creative arts, and social behaviour.
  
• Interventional Designs: Could AI tools be designed to stimulate engagement and cognitive presence (e.g., adaptive prompting systems), rather than induce passivity? Early AI should augment, not replace, mental effort.
  
• Neuroimaging Studies: EEG, fNIRS, and fMRI could offer high-resolution insight into brain dynamics during AI multigenerational interactions.
  
• Ethical Considerations: With AI’s growing intimacy in our lives and even direct neural coupling, we must consider consent, mental autonomy, emotional impact, and cognitive health policy.

Conclusion

This research frontier is growing, and there’s increasing evidence that while AI offers powerful cognitive tools, excessive reliance may reduce brain engagement and cognitive skills like memory, creativity, and critical thinking. Encouragingly, the adult brain remains plastic, and AI could be designed—ethically and thoughtfully—to support rather than erode our cognitive capacities.