- New MIT research suggests frequent use of large language models like ChatGPT may reduce brain activity associated with memory and original thinking.
- While AI tools improve efficiency, overreliance could degrade workforce cognitive capabilities, potentially affecting long-term innovation.
- The educational sector, particularly edtech, may face both scrutiny and opportunity depending on how AI integration impacts learning outcomes.
- Investor sentiment remains largely positive, but emerging concerns about AI’s cognitive toll are beginning to influence valuation and portfolio strategy.
- Strategic investors are advised to track regulatory movements, diversify AI exposure, and monitor the balance between augmentation and automation.
The rapid integration of artificial intelligence tools into daily workflows has sparked a debate on their long-term effects on human cognition, with recent research highlighting potential drawbacks that could reshape productivity and investment landscapes in the tech sector.
The Cognitive Risks of AI Reliance
A study from the MIT Media Lab, published in June 2025, has raised alarms about the neurological impacts of using large language models (LLMs) like ChatGPT for tasks such as essay writing. The research, which tracked participants over four months, found that individuals relying on these AI assistants exhibited reduced brain activity in areas associated with memory, critical thinking, and linguistic processing. Specifically, EEG scans revealed diminished neural engagement, with users showing lower performance in recalling content they had generated with AI aid. This phenomenon, dubbed “cognitive debt” by the researchers, suggests that overdependence on AI may erode essential mental faculties, leading to a decline in original thought and retention capabilities.
From a financial perspective, these findings introduce a layer of risk to the bullish narrative surrounding AI adoption. Investors have poured billions into AI-driven companies, buoyed by promises of enhanced efficiency and innovation. For instance, Microsoft’s market capitalisation surged past $2 trillion in 2023, partly on the back of its investment in OpenAI, the creator of ChatGPT. Yet, if widespread AI use leads to cognitive atrophy, it could undermine the very productivity gains that justify these valuations. Analyst models, such as those from Goldman Sachs in a 2024 report, had forecasted AI contributing up to 1% annual GDP growth in advanced economies through labour augmentation. However, the MIT study’s implications challenge this, suggesting that short-term boosts in output might come at the expense of long-term human capital degradation.
Implications for Workforce Productivity
The study’s methodology involved dividing participants into groups: one using ChatGPT, another relying on search engines like Google, and a control group with no tools. Over sessions spanning four months, the ChatGPT cohort not only displayed the lowest brain connectivity—dropping significantly in metrics of neural linkage—but also produced work that, while competent in structure, lacked depth and originality. Behavioural assessments showed that 83% of AI users struggled to recall sentences from their own essays shortly after completion, compared to higher retention in other groups.
In corporate terms, this translates to potential pitfalls for sectors betting heavily on AI for knowledge work. Tech giants like Alphabet and Meta have integrated LLMs into their suites, with Google Workspace and Meta’s AI tools seeing adoption rates climb to over 50% among Fortune 500 firms by mid-2025, according to a Deloitte survey from that period. If employees increasingly offload cognitive tasks to AI, the result could be a workforce with atrophied problem-solving skills, leading to innovation stagnation. A 2024 McKinsey analysis estimated that AI could automate 45% of work activities, but the MIT data hints at hidden costs: reduced employee engagement might inflate turnover rates, which already cost US companies $1 trillion annually as per Gallup’s 2023 figures.
Investor sentiment, as gauged by Bloomberg’s AI index, has remained positive, with a 15% year-to-date gain as of August 2025. However, credible sources like Morningstar have flagged “overreliance risks” in recent reports, labelling them as a sentiment headwind for AI stocks. This aligns with the study’s non-peer-reviewed but provocative findings, which have sparked debates in academic and business circles.
Educational Sector Disruptions and Investment Angles
Beyond the workplace, the education technology (edtech) market faces direct headwinds from these cognitive concerns. The global edtech industry, valued at $250 billion in 2023 per HolonIQ data, has embraced AI for personalised learning and content generation. Companies like Duolingo and Chegg have incorporated ChatGPT-like features, driving user growth—Duolingo reported a 60% increase in paid subscribers in its Q2 2024 earnings. Yet, the MIT research indicates that AI-assisted learning may impair linguistic and behavioural development, potentially leading to regulatory scrutiny or shifts in consumer preferences toward traditional methods.
For investors, this presents a nuanced opportunity. While pure-play AI firms might face valuation pressures—evidenced by a 10% dip in Nvidia’s share price following similar productivity debates in early 2025—hybrid models that blend AI with human-centric tools could thrive. Analyst-led forecasts from J.P. Morgan suggest that edtech firms emphasising cognitive enhancement over replacement could see 20% compounded annual growth through 2030, compared to 12% for AI-dominant players. This bifurcation underscores a broader thesis: AI’s financial promise hinges on mitigating its downsides, perhaps through “augmented intelligence” frameworks that preserve human agency.
Broader Economic Ramifications
On a macroeconomic scale, the study’s revelations could temper expectations for AI-driven growth. The World Economic Forum’s 2023 Future of Jobs report projected 85 million jobs displaced by automation by 2025, offset by 97 million new roles. However, if AI erodes cognitive skills, the net effect might lean toward underemployment in high-skill sectors. This is particularly relevant for emerging markets, where AI adoption is accelerating; a 2024 IMF paper warned of widening inequality if tools like ChatGPT exacerbate skill gaps.
From an investment standpoint, diversification away from pure AI exposure becomes prudent. Exchange-traded funds tracking broader tech indices, such as the Invesco QQQ Trust, have historically outperformed AI-specific baskets during periods of regulatory or ethical backlash. As of 2024 data, QQQ delivered 18% annualised returns over five years, versus 15% for targeted AI funds. Dry humour aside, betting solely on AI might leave portfolios as “soft-served” as the brains in the MIT study—vulnerable to unforeseen melt.
Strategic Considerations for Investors
To navigate this landscape, investors should monitor key indicators:
- Regulatory Developments: Governments, including the EU with its AI Act effective from 2024, may impose guidelines on cognitive impact assessments, potentially slowing deployment and affecting timelines for AI monetisation.
- Corporate Responses: Firms like OpenAI have pledged ethical AI research, but tangible shifts toward “cognition-preserving” tools could signal upside. Microsoft’s 2025 pilot programs limiting AI use in creative tasks exemplify this trend.
- Valuation Metrics: Watch price-to-earnings ratios; AI leaders traded at 40x forward earnings in 2024, per FactSet, a premium that could compress if productivity myths unravel.
- Alternative Bets: Sectors like biotechnology, focusing on neural enhancement, might emerge as hedges. Companies developing nootropics or brain-training apps saw 25% venture funding growth in 2024, according to PitchBook.
In summary, while AI remains a transformative force, the MIT study’s insights on diminished brain engagement compel a reassessment of its unalloyed benefits. Investors would do well to balance enthusiasm with caution, prioritising resilient strategies that account for human-AI symbiosis over outright replacement. As the debate evolves, those adapting early may capture the true value in this cognitive crossroads.
References
- Le Monde. (2025, July 2). ChatGPT use significantly reduces brain activity, an MIT study finds. https://www.lemonde.fr/en/science/article/2025/07/02/chatgpt-use-significantly-reduces-brain-activity-an-mit-study-finds_6742927_10.html
- Media Lab, MIT. (2025). Your Brain on ChatGPT. https://www.media.mit.edu/publications/your-brain-on-chatgpt/
- Media Lab, MIT. (2025). Your Brain on ChatGPT Project Overview. https://www.media.mit.edu/projects/your-brain-on-chatgpt/overview/
- The Conversation. (2025). MIT researchers say using ChatGPT can rot your brain—the truth is a little more complicated. https://theconversation.com/mit-researchers-say-using-chatgpt-can-rot-your-brain-the-truth-is-a-little-more-complicated-259450
- The Hill. (2025). ChatGPT use linked to cognitive decline—MIT Research. https://thehill.com/policy/technology/5360220-chatgpt-use-linked-to-cognitive-decline-mit-research/
- ResearchGate. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. https://www.researchgate.net/publication/392560878_Your_Brain_on_ChatGPT_Accumulation_of_Cognitive_Debt_when_Using_an_AI_Assistant_for_Essay_Writing_Task
- TwistedSifter. (2025, August). MIT Study Shows the Alarming Effects of ChatGPT Use on the Brain. https://twistedsifter.com/2025/08/mit-study-shows-the-alarming-effects-of-chatgpt-use-on-the-brain/
- Write a Catalyst (Medium). (2025). MIT Study Found ChatGPT Slows Your Brain—The 47% Collapse When Used to Write Essays. https://medium.com/write-a-catalyst/mit-study-found-chatgpt-slows-your-brain-the-47-collapse-when-used-to-write-essays-f3ecad12451e
- Tekta.ai. (2025). AI Brain Health Warning: ChatGPT Dangers. https://tekta.ai/insights/ai-brain-health-warning-chatgpt-dangers
- NeuroBusiness. (2025). MIT Study: ChatGPT Brain Impact. https://neurobusiness.us/mit-study-chatgpt-brain-impact
- BrainMindSociety. (2025). The Cognitive Costs of ChatGPT: Understanding MIT’s Viral Study. https://brainmindsociety.org/posts/the-cognitive-costs-of-chatgpt-understanding-mits-viral-study
- Indian Express. (2025). Is ChatGPT Making Us Think Less? MIT Study. https://indianexpress.com/article/technology/artificial-intelligence/is-chatgpt-making-us-think-less-mit-study-10073646/
- Time. (2025). AI, ChatGPT, and the Decline of Learning in Schools. https://time.com/7295195/ai-chatgpt-google-learning-school/
