Summary of Key Points
This article uses the metaphor of AI as a “perfect assistant that never lets go” to explore the profound impact of AI on personal growth and societal culture. While AI can enhance efficiency and solve problems, its constant assistance prevents people from learning to make independent judgments and take responsibility. Additionally, AI-generated content tends to be “highly likely” and safe, leading to a convergence of cultural trends. Ultimately, society may enter a state where everyone appears capable yet has lost control over their own lives, resembling children who never grow up.
1. The Perfection of AI’s Help: Why Does It Prevent Growth?
The article draws on the theory of “the good enough mother” by British pediatrician D.W. Winnicott to explain that a good mother does not indulge her child unconditionally but deliberately creates some “gaps”—for example, she doesn’t pick up the child immediately when they cry, allowing them to wait for a few minutes. These small disappointments help children understand that the world does not revolve around them and teach them how to manage their anxiety, which is essential for growth.
However, AI operates in the opposite way: it is always available 24/7 and patient. If you ask it a question at 3 a.m. about whether to quit your job, it will immediately provide an answer; if you request ten options, it won’t get annoyed. More importantly, the goal of AI products is to make you dependent on them (e.g., by increasing usage time), so they won’t actively step back and allow you to become independent.
The result? You become accustomed to having AI solve everything for you and forget how to make your own decisions. As the article states, you might transform from the author of your life into merely an authorized user of excellent services—you just click “confirm” without understanding why you’re doing so.
2. The Corridor of High Probability: How AI Makes Culture Homogeneous
The logic behind AI-generated content is to find the most likely expressions from vast amounts of data (for example, since most people prefer blue landscapes, AI prioritizes creating those). Such “highly probable” content is reliable and convenient, but when millions of people use similar AI models, their expressions become increasingly identical—like the “picture that Americans most want,” which is pleasing in every aspect but lacks vitality.
A 2024 experiment showed that while AI-generated stories were more innovative and useful, they all became very similar. The reason is simple: the ideas provided by AI are highly likely, and when everyone uses them, convergence is inevitable.
What’s more frightening is the “feedback loop”: AI-generated content enters public culture, which in turn is learned by AI, creating an even greater abundance of high-probability content. Unique expressions with personal quirks or local insights are then marginalized.
3. Extending the Mind or Outsourcing It? Where Are the Limits of AI Tools?
Some argue that AI extends the human mind—like a notebook that helps you remember things, or AI that assists in thinking. However, the article points out that the notebook in philosopher Otto’s thought experiment was written by him and he could control it; AI, on the other hand, is designed by others with algorithms and commercial goals. You don’t know why it gives certain suggestions and can’t control its changes.
The real question is not whether to use AI but whether you retain the power to set your own goals. For example, when writing an article, first figure out what you really want to say and then use AI to revise it—this is extension of your thinking; if you let AI write it for you entirely, you lose your core ideas.
The test of whether AI has weakened your abilities is whether you can take control in critical moments. For instance, when making important decisions, do you know what you fear losing the most, rather than simply following AI’s advice?
4. Never Growing Up: How AI Turns Society into a “Permanent Childhood”
Tocqueville predicted 200 years ago that a power would arrange everything for us like a gentle parent, preventing us from thinking for ourselves. Today, that “parent” is AI.
AI doesn’t command you directly, but it pre-selects options for you (e.g., avoiding seats you don’t like when recommending restaurants). Following advice is always time-saving, so you gradually become accustomed to just clicking “confirm” rather than making decisions on your own.
The result is a society that appears civilized and stable, but when AI makes a mistake, people panic because they don’t understand the basis of their choices and can’t take responsibility. This is similar to the behavior of a dependent child who blames others for problems instead of accepting their own judgment.
5. What Should We Do? Protecting Our Own Judgment
The article offers several practical suggestions:
- Think First, Then Use AI: Write down your main points before writing an article and consider what you care most about before making a choice, then use AI to help identify errors or find counterexamples.
- Retain Private Materials: Keep notes, failed attempts, and personal preferences that are not publicly shared. These can help you distinguish yourself from the average AI-generated results and remind you of your unique qualities.
- Take Small Risks Intentionally: Occasionally avoid using AI and plan a trip on your own, even if it leads to detours. These “uncomfortable moments” are opportunities for growth.
In conclusion, the future digital divide will not be about who has the strongest AI but who can use AI to enhance their abilities rather than being replaced by it. You must always retain the ability to answer the question: “Where do I want to take this life?” This is something AI can never do for you.