The Age of Brain Outsourcing Has Arrived
Although the concept of brain outsourcing is not new and has been discussed in earlier writings, the arrival of hard scientific evidence makes its consequences all the more alarming. When we use tools like ChatGPT to draft emails, generate reports, or brainstorm ideas, are we really assisting our thinking—or outsourcing it?
Recent research by the Massachusetts Institute of Technology (MIT) has provided critical insights into this very question. The study confirms what many may have intuitively suspected: that heavy reliance on AI tools may be compromising our brain’s ability to think and remember.
The MIT Study: Three Groups, One Outcome
In MIT’s four-month-long experiment, 54 participants were divided into three groups:
1. ChatGPT group – who used ChatGPT to assist in writing
2. Google Search group – who used only traditional search engines
3. Brain-only group – who relied entirely on their own thinking and writing
All participants were equipped with EEG devices to monitor brain activity while completing multiple writing tasks.
The results were striking: those who heavily relied on ChatGPT exhibited a significant decrease in neural connectivity. When asked to write independently without AI, their brain activity resembled that of a novice, not an experienced writer. This is a direct symptom of brain outsourcing—a cognitive version of hiring a chauffeur so frequently that you forget how to drive.
Cognitive Forgetfulness: A Silent Crisis
One particularly unsettling finding was the emergence of cognitive forgetfulness. A staggering 83% of ChatGPT users reportedly failed to accurately recall sentences they had written—just minutes earlier—using AI assistance. In comparison, only 11% of the brain-only group experienced this issue.
(Note: These exact figures—83% vs 11%—come from summaries of the preprint study. The paper has not yet been peer reviewed and is still undergoing formal evaluation.)
This disparity reflects how outsourcing cognitive processes to AI leads us to become information couriers, rather than knowledge internalizers.
MIT researchers coined the term “cognitive debt” to describe this phenomenon: trading away future cognitive capacity in exchange for immediate convenience. The analogy to GPS and our fading sense of direction is apt—but this time, the damage is happening at the level of thought itself.
Ancient Wisdom, Modern Validation
Interestingly, this modern scientific insight aligns closely with ancient wisdom, particularly from Japanese philosophy. The traditional concept of Shu-Ha-Ri (守破離)—a three-stage learning model rooted in martial arts and tea ceremony—offers a profound perspective on the AI age:
Shu (守): The apprentice faithfully follows the master's instructions and internalizes the basics
Ha (破): Once mastery is achieved, the learner begins to break away and seek their own path
Ri (離): Finally, the practitioner transcends tradition and innovates independently
Using AI to generate content without first mastering the basics is akin to skipping “Shu” and jumping straight to “Ha” and “Ri”. The results may look polished but lack true depth.
Similarly, Japanese craftsmanship (職人精神) teaches us that excellence stems from relentless repetition and attention to foundational skills. A sushi apprentice may spend years just learning how to cook rice—a reflection of the belief that no shortcut leads to mastery.
The Neuroscience Behind It All
Modern neuroscience backs this up through the principle of neuroplasticity—the “use it or lose it” nature of our brains. Challenging cognitive tasks strengthen neural networks, while outsourcing them weakens those same pathways.
True learning occurs when we engage in deliberate practice and embrace desirable difficulty—conditions that demand effort and foster deep, lasting memory. The seamless ease offered by AI removes these very challenges, replacing deep learning with superficial familiarity.
Coexisting with AI: A Smarter Path Forward
So how should we coexist with AI tools like ChatGPT? The answer is not to ban them, but to use them wisely.
The MIT study revealed a hopeful insight: participants who first wrote a draft independently, and only then used ChatGPT for editing or enhancement, actually showed increased neural connectivity.
This suggests a healthier approach: bear the cognitive burden yourself first (Shu), then use AI as a refinement tool (Ha, Ri).
This is especially vital for the younger generation and beginners in any field. AI should be seen as a powerful collaborator, not a replacement for human thinking. Delegating all cognitive work to AI may seem efficient—but it comes at the cost of weakening the very faculties that make us human.
Final Thoughts
The path to human-AI synergy begins not with full reliance, but with intentional partnership. Let’s not be so eager to enjoy the benefits of technology that we unknowingly accumulate an unpayable cognitive debt.
Don’t outsource your brain before you've even begun to use it.
—
Dr. Ken FONG
中文摘要
大腦外包已經成為事實，而科學研究揭示了它的代價。
MIT 近期的研究針對 54 位參與者進行為期四個月的追蹤，透過腦電圖 (EEG) 分析他們在不同寫作方式下的大腦活動。結果顯示：長期依賴 ChatGPT 協助寫作的用戶，其大腦連結性顯著下降，與新手寫作者無異。這種現象被稱為「大腦外包」。
更令人擔憂的是「認知失憶」：高達 83% 的 ChatGPT 使用者無法記得自己幾分鐘前寫過的句子，因為他們並未真正參與思考與組織過程，淪為資訊的搬運工。相對之下，「純大腦組」只有 11% 遇到同樣問題。
（註：83% 與 11% 這兩個數字來自於該研究的前期摘要與媒體報導。該論文目前尚未通過正式的同行評審，仍處於預印本（preprint）階段，後續仍需經過正式審查與發表。）
研究團隊將這種現象命名為「認知債務 (cognitive debt)」：使用未來的思考能力來換取眼前方便。
日本哲學「守破離」與職人精神也是文化精點，也是中能體現實質的教訓。相信執著基礎不厭的技能磨練，才是真正的優秀來源。
現代神經科學說明：大腦可塑性 (neuroplasticity) 依賴於「用進廢退」，而「計劃練習」與「有益困難」是加強記憶、深度學習最有效的方式。而 AI 給的方便性則總算是換來的代價。
因此，真正的合理存在方式是：先成功承擔認知重擔，再利用 AI 進行加強與優化。對於學習者與新人者來說，都應該把 AI 視為合作夥伴，而不是代理者。
小心，別一開始就把你的思考給外包了。
References
- MIT EEG Research on Cognitive Impact of AI-assisted Writing
- Principles of Neuroplasticity and Desirable Difficulty
- Japanese Philosophy: Shu-Ha-Ri and Shokunin Spirit
Keywords
brain outsourcing, cognitive debt, MIT study, ChatGPT impact, neuroplasticity, cognitive forgetfulness, desirable difficulty, AI-assisted writing, Shu-Ha-Ri, deliberate practice, human-AI collaboration, deep learning, digital tools, productivity and cognition, generative AI ethics, knowledge internalization, cognitive decline, educational psychology, smart AI usage, professional thinking
Posts
0 Comments