The phenomenon of brain outsourcing is beginning to spread among people who heavily use AI, and experts explain that this is extremely dangerous. (Background: How is AI changing human reading habits? Will original texts disappear?) (Additional background: A Berkeley professor warns: Graduates from prestigious schools also have no job choices! AI will cut half of entry-level positions in 5 years) Technology always comes from inertia. Please honestly reflect: have you, who use AI daily for work, writing, or even creation, become a part of 'brain outsourcing'? Perhaps you haven't yet understood what brain outsourcing is. It differs from the habit of 'externalizing memory' cultivated over decades by the common use of search engines like Google. Brain outsourcing structurally undermines human originality and creativity, pushing people's minds more towards mechanical goals rather than planning and execution. Experts warn that the founder of the well-known domestic AI company iKala, Sega Cheng, recently wrote a shocking post on Facebook, which has been authorized to share in full below: Brain outsourcing has fully launched. Although I mentioned this in my book a long time ago, I was still quite surprised by the severity of this issue once the real scientific research results emerged. When we input commands in the dialogue box to let ChatGPT draft an email for us, write a report, or even brainstorm ideas, are we assisting our thinking or unconsciously engaging in a large-scale 'brain outsourcing'? Last week, MIT provided us with scientific evidence. Although the research results may have been anticipated by most people, they were not particularly surprising. Researchers recruited 54 participants, equipped them with EEG (electroencephalogram) devices, and asked them to complete multiple articles over a period of 4 months. They were divided into three groups: ChatGPT group: using ChatGPT to assist in writing. Google search group: only using traditional search engines. Pure brain group: relying solely on their own brain for thinking and writing. The results showed that long-term users of ChatGPT had significantly decreased brain connectivity. When they were asked to write independently without AI, their brain activity patterns resembled those of novice writers rather than experienced authors. This is the direct consequence of brain outsourcing: you completely hand over the cognitive burden of constructing arguments and organizing text to AI, just like hiring a driver and forgetting how to drive. Even more unsettling is a phenomenon of 'cognitive amnesia': as high as 83% of ChatGPT users were unable to accurately quote sentences they had written themselves (with AI assistance) just minutes earlier. Because the process of thinking and organizing has been outsourced, we have become 'movers' of information rather than 'internalizers' of knowledge. In contrast, only 11% of the pure brain group faced the same difficulty. This is a tremendous gap: 83% vs 11%. Researchers named this consequence 'cognitive debt': you trade your future cognitive abilities for short-term convenience. This is similar to how GPS has created a large group of people who can't navigate; excessive reliance on GPS weakens our inherent sense of space and ability to find our way. Now, ChatGPT has elevated the level of outsourcing from 'direction sense' to 'thinking' itself. This is very concerning, after all, 'thinking ability' is key to our self-proclaimed status as the most intelligent beings. This scientific discovery actually aligns with many ancient wisdoms (science often helps verify ancient wisdom). Before computers appeared, many cultures warned us of the dangers of taking shortcuts, most profoundly expressed in Japanese philosophy: 'Shuhari' is a three-stage learning theory derived from traditional arts like martial arts and tea ceremonies, perfectly illustrating the challenges of the AI era. Shuhari: Protect, adhere. Apprentices must completely imitate and faithfully follow the teachings and all basic skills of their masters, without personal interpretation, aiming to internalize the basics as instincts. Break: Break through, shatter. After mastering the basics, apprentices begin to question traditions, seeking methods that better suit themselves. Leave: Depart, transcend. Ultimately, learners fully surpass the original forms and create their own new paths. Directly using AI to generate content is like an apprentice who has never even experienced the 'Protect' stage wanting to jump straight into the realms of 'Break' and 'Leave.' The results may appear splendid, but the foundation is unstable. Similarly, Japan's 'artisan spirit' also values achieving mastery in a skill. A sushi chef's apprentice spends years just to learn how to cook rice, embodying this spirit. True excellence comes from relentless honing of the basics. This cultural wisdom of 'not taking shortcuts' is not mere philosophical talk; it is rooted in the scientific principles of how the brain operates. First is neuroplasticity, which is the brain's characteristic of 'use it or lose it.' When we strive to think and solve problems, the relevant neural connections are strengthened. Conversely, long-term outsourcing of these tasks weakens connectivity. Taking shortcuts will come at a price later. If you take shortcuts in thinking, the cost you pay is poorer neural connectivity in your brain later on, and your mind will become less effective. The best way to strengthen brain connectivity comes from deliberate practice and desirable difficulty. Our brains learn best and form the deepest long-term memories when faced with challenges that are beyond our comfort zones and of certain difficulty. The seamless convenience provided by AI strips us of the opportunity to experience this 'beneficial difficulty,' causing knowledge to float on the surface without taking root. So how should we coexist with AI? The answer is certainly not to ban it, but to use it wisely. Another key finding from MIT's research gives us a glimmer of hope: participants who first completed initial drafts independently and then used ChatGPT for refinement and editing showed an increase in brain connectivity. Therefore, the whole conclusion is very clear: first use your own brain to tackle the cognitive burden (Protect), then use AI as a reinforcement tool for breakthroughs (Break, Leave). This conclusion is crucial for the next generation of children and beginners in all fields. We should view AI as an immensely potential 'collaborator,' rather than a 'substitute' to outsource our thinking, tossing tasks to AI without a second thought. Only then can we enjoy the benefits of technology while avoiding incurring irredeemable 'cognitive debt,' ensuring we remain truly capable of independent thought, and not outsourcing our brains from the very beginning. Don't give up thinking from the start. The severity of brain outsourcing can be described with a term that isn't a real medical term: 'cognitive laziness syndrome.' When cognitive tasks become too easy, the brain reduces its engagement, thereby weakening the activity of neural networks, ultimately leading to a tangible decline in cognitive function. For example, in stroke patients, if rehabilitation is not pursued, the areas of the brain that are not used will simply be sealed off. Or to use a more extreme analogy: if one sits in a wheelchair for a year, can those legs still walk? In the past two years, behavioral scholars have also begun to refocus on the research related to human neuroplasticity due to the prevalence of AI. Each time people choose to let AI do the thinking for us, we are actually missing an opportunity for learning and reinforcement...

#AI