AI Is 'Eating' Your Brain

AI vs. Attention

The stark claim: AI is not just automating tasks — it is eating our attention and rewiring how we think. Reportedly, a 2025 preprint from MIT Media Lab using EEG data found that long‑term reliance on generative AI for writing was associated with markedly lower neural activation and 45–55% fewer active brain‑region connections when participants attempted unaided tasks. The study is a preprint with 54 participants from the Boston area and has not been peer‑reviewed; its authors cautioned that this is not evidence of physical brain atrophy, but of changing neural activity patterns.

Neuroscience and behavioural economics help explain the worry. Herbert A. Simon’s old insight — more information means less attention — still holds. Platforms like TikTok (抖音海外版), run by ByteDance (字节跳动), by design exploit dopamine and variable rewards; generative AI now promises richer, subtler ways to fragment attention by pre‑empting thought rather than merely interrupting it. How do you defend your focus when the “assistant” proposes the frame before you even form a question?

Economic and geopolitical fallout

The consequences are social as well as cognitive. A 2023 Goldman Sachs estimate suggested AI could affect some 300 million full‑time jobs globally. Industry reports, including an Anthropic analysis released this March, estimate high automation potential across programming, customer service and data tasks. Meanwhile, compute and talent are concentrated in a handful of firms — OpenAI, Google, Microsoft, and Chinese players such as Baidu (百度) and Alibaba (阿里巴巴) — and access to high‑end chips is shaped by geopolitics and export controls. The result: an enormous creation of value that most people may not meaningfully share.

What to do about it

That leaves a practical question: how do individuals and societies respond? Experts interviewed in China’s tech press and elsewhere suggest a three‑step approach: use human cognition first to frame problems; use AI to expand options and draft; and finally use independent, deep reflection to integrate and judge AI output. It sounds simple. But in a world where attention is scarce and technological advantages are asymmetric, reclaiming the mental space to do that will be as much a political and educational challenge as a personal one.