Research scientist Nataliya Kosmyna and her colleagues at the MIT Media Lab recruited 54 students to write short essays, dividing them into three groups. The first group was asked to use no technology at all. The second group could use Google Search, but with AI-generated summary features turned off. The third group used ChatGPT. Researchers measured brainwaves while each student was writing. Was there a difference in brain activity across the groups?
🔖 Preliminary Research Findings
1️⃣Group 1 (No Technology): Brains were exceptionally active, showing widespread activity across multiple regions.
2️⃣Group 2 (Google Search): Brains were moderately active, with strong activity remaining in visual processing regions.
3️⃣Group 3 (ChatGPT): Brain activity significantly decreased—up to 55% lower than Group 1. Activity in regions related to creativity and information processing was particularly diminished.
🔖 Potential Long-term Impacts
⚠️ Loss of Professional Judgment: Known as “Cognitive Surrender,” individuals tend to accept AI-provided information almost without question, even allowing it to override their own judgment.
⚠️ Risk of Cognitive Decline: Due to the lack of deep thinking and information integration, this may lead to a decline in cognitive ability and spatial memory, increasing the risk of premature brain aging.
🔖 Other Key Discoveries
✅”No Pain, No Gain”: Those who thought for themselves and built a solid foundation of ideas before using AI showed much higher brain activity than those who used AI from the very start.
✅”Human-in-Command”: Those who treated AI as a tool for data collection but conducted the final analysis themselves generally made more accurate predictions and showed stronger brain activation.
As AI becomes increasingly convenient and powerful, are we using it to “evolve,” or are we letting our brains “retire early”?
Reference:
BBC News https://www.bbc.com/future/article/20260417-ai-chatbots-could-be-making-you-stupider
研究科學家Nataliya Kosmyna和麻省理工學院媒體實驗室的同事招募了54名學生撰寫短文，並將他們分成三組。第一組被要求不使用任何科技產品。第二組可以使用Google搜索，但關閉了AI生成的摘要功能。第三組則使用ChatGPT。研究人員在每位學生寫作時都測量了他們的腦電波。到底每一組的學生的腦電波有否分別？
🔖初步研究結果
1️⃣第一組（不使用任何科技產品）：大腦異常活躍，大腦多個區域都表現出廣泛的活動。
2️⃣第二組（使用Google搜索）：大腦中度活躍，大腦視覺區域仍然表現出強烈的活動。
3️⃣第三組（使用ChatGPT）：大腦活動明顯減少，比第一組減少了高達 55%，特別是與創造力和資訊處理相關的區域的活躍度大大降低。
🔖可能的長期影響
⚠️喪失經年訓練的專業判斷能力：認知投降，傾向於幾乎不加質疑地接受人工智慧給予的訊息，甚至允許它凌駕於自己的判斷之上。
⚠️大腦功能退化風險：由於缺乏深層思考與資訊整合過程，可能導致認知能力、空間記憶力下降，增加大腦早衰風險。
🔖其他研究發現
✅「先苦後甜」：那些先靠自己思考、打好思考基礎後再使用 AI 的人，腦部表現比「一開始就用 AI」的人活躍得多。
✅「以我為尊」：那些將AI作為工具來收集數據，然後自行進行分析的人普遍做出了更準確的預測，並且表現出更強的腦部活化。
參考資料:
BBC News https://www.bbc.com/future/article/20260417-ai-chatbots-could-be-making-you-stupider