论人工智能解放创造力

On the Liberation of Creativity by Artificial Intelligence

在2025年，人类毫无疑问已经一只脚迈进了人工智能的时代。OpenAI的语言大模型ChatGPT在2022年底的兴起，以前所未有的速度将人工智能普及到了人们的日常工作与生活之中。在那之后，OpenAI、Google、Anthropic、Meta等科技巨头开始大力压注人工智能赛道，大模型之间的竞争很快变得白热化。仅仅两年多之后，大模型的能力就得到了极大的进化，目前最强大的大模型（Gemini 2.5 Pro, Grok 4, GPT-o3等）在文字生成、数学、代码生成、图像生成等方面普遍已经达到了等价于硕士研究生的水准 —— 即使是一年前这都是难以想象的。

In 2025, humanity has undoubtedly stepped into the era of artificial intelligence with one foot already through the door. The rise of OpenAI's large language model, ChatGPT, at the end of 2022, popularized artificial intelligence in people's daily work and lives at an unprecedented speed. Following that, tech giants like OpenAI, Google, Anthropic, and Meta began to invest heavily in the AI race, and the competition between large models quickly became white-hot. Just over two years later, the capabilities of these models have evolved tremendously. The most powerful large models today (such as Gemini 2.5 Pro, Grok 4, GPT-o3, etc.) have generally reached the level equivalent to a master student in areas like text generation, mathematics, code generation, and image generation—a feat that was hard to imagine even a year ago.

现在这个时间点，对于人工智能未来的发展上限，社会中充满了激烈的争论。包括我个人在内的一部分人认为人工智能的能力并没有上限，可能在10年之内就会达到舆论场所谓的"通用人工智能"甚至"超级人工智能"的水平；但也有相当一部分人不相信人工智能能够在一些所谓的"高阶思维能力"之上真正达到人类的水平。但无论持哪一方观点，一个已经基本形成的共识是：人工智能很快将会能够执行那些结构化的、重复性的、更多依靠记忆和经验的工作。对于这一变化所造成的社会影响，很多人工智能领域的舆论领袖都给出了一个乐观的、也似乎是符合工业革命以来的历史演进方向的预测：人类将会从繁重而无趣的机械性劳动中解放出来，转向更高价值的创造活动。这似乎也是这一轮人工智能革命能够为人类带来的一个重大的福音。

At this point in time, society is filled with intense debate about the upper limits of artificial intelligence's future development. One group of people, myself included, believes that there is no ceiling to AI's capabilities and that it may reach the level of what is publicly termed "Artificial General Intelligence" (AGI) or even "Artificial Superintelligence" (ASI) within ten years. However, a considerable number of people do not believe that AI can truly match human levels in so-called "higher-order thinking abilities." But regardless of which viewpoint one holds, a consensus has largely formed: artificial intelligence will soon be able to perform tasks that are structured, repetitive, and rely more on memory and experience. Regarding the societal impact of this change, many thought leaders in the AI field have offered an optimistic prediction, one that also seems to align with the historical trajectory since the Industrial Revolution: humanity will be liberated from tedious and uninteresting mechanical labor and shift toward higher-value creative activities. This, it would seem, is a major blessing that this AI revolution can bring to humankind.

对于这一判断本身，我是基本认可的。虽然我不认为创造力是人类所独有的能力，但至少目前的人工智能的创造力的确还相当低级，并没有真正学到少数人类（如爱因斯坦、乔布斯等）所拥有的那种基于"第一性原理"的原创能力，因此至少这一趋势在短期的未来应当是成立的。这里我想讨论的恰恰是一个大众常常都忽略的关注点：从繁重而无趣的机械性劳动转向更高价值的创造，这一判断在情感上并不是中立的，而是带有一种显然的善的倾向。从现代社会的舆论宣传、以及多数人自身经验中，很多人的确会认为那种重复性的劳动会折磨人的心智，而创造性活动则更可能为人带来愉悦的体验。因此，当舆论领袖们在作出这样的乐观预言时，只要认可其背后的逻辑链条，听众在情感上也都会认同其积极的一面。我想提出的疑问便是：事情真的有我们直觉上体会到的那么简单吗？

Basically, I agree with this judgment itself. While I don't believe that creativity is an exclusively human ability, the creative capacity of current AI is, at least for now, indeed quite rudimentary. It hasn't truly mastered the kind of originality, rooted in 'first principles,' that a select few humans (such as Einstein or Steve Jobs) possess. Therefore, this trend should hold true at least in the near future. What I want to discuss here is precisely an aspect that is often overlooked by the public: the judgment that we will shift from arduous and tedious mechanical labor to higher-value creative work is not emotionally neutral; rather, it carries a clear bias towards what is considered 'good.' Based on public discourse in modern society and the personal experiences of most people, many indeed believe that repetitive labor can be soul-crushing, whereas creative activities are more likely to provide a sense of joy and fulfillment. Consequently, when opinion leaders make such optimistic predictions, as long as the audience accepts the underlying logic, they will also emotionally resonate with its positive aspects. The question I wish to pose is this: Is the matter truly as simple as our intuition suggests?

为了更深入地理解这个问题，我们不妨做一个思想实验：考虑一个所有结构化的重复性劳动都已被人工智能所替代的社会，人类之间的市场竞争转向了高价值的创造力之间的竞争。在这个社会之中，由于互联网高度发达、人工智能的执行力无比强大，因此一个好的创意能够很快被人工智能转化为实际的产品，并且通过互联网与全球供应链迅速地渗透到全球市场。那么这样的一个假想的市场会有怎样的特点？答案是相当直接而残酷的：这会是一个极度"赢家通吃"的市场。事实上，对于高度依赖创意的产业的从业者而言，这应当是一个近乎常识的认知。那些需要重复性劳动的领域，虽然工作可能劳累而又无趣，但它往往是"一分耕耘一分收获"，投入多少精力就会带来多少回报；创意行业则完全不同，一个天才的想法就可以让一千个平凡的想法失去价值。如今的创造性工作至少还在很大程度上受到劳动效率的限制，因此可以在一定程度上压制赢家通吃效应；可是当劳动效率的瓶颈完全被人工智能攻破了呢？那整个世界真正能贡献价值的人可能永远都不会超过千分之一的比例。

To delve deeper into this issue, let's conduct a thought experiment: consider a society where all structured, repetitive labor has been replaced by artificial intelligence, and market competition among humans has shifted to a contest of high-value creativity. In such a society, thanks to a highly advanced internet and the immense executive power of AI, a good idea can be rapidly transformed into an actual product and, through the internet and global supply chains, quickly penetrate the global market. What, then, would be the characteristics of such a hypothetical market? The answer is rather straightforward and brutal: it would be an extreme "winner-take-all" market. In fact, for practitioners in industries that heavily rely on creativity, this should be an almost self-evident realization. In fields requiring repetitive labor, while the work may be exhausting and dull, it is often a case of "you reap what you sow"—the effort you invest yields a proportional reward. The creative industries, however, are entirely different; a single stroke of genius can render a thousand mediocre ideas valueless. Currently, creative work is at least still constrained to a large extent by the limits of labor efficiency, which helps to suppress the winner-take-all effect to some degree. But what happens when this bottleneck of labor efficiency is completely shattered by AI? In that case, the proportion of people in the entire world who can truly contribute value may never exceed one-thousandth.

这就是创造力竞争结合市场机制所形成的图景，同时也是我们这个社会在不断地神化创造力这一品质时，常常有意忽略的一个沉重现实。创造本身当然是美好的，但必须接受市场检验的创造则是极为残酷的。我们用于指引自身的杰出人物通常都有强大的创造力；但更大比例的拥有同等级甚至更高创造力的人杰，因为各种原因没有能够让他们的创造被世人所接受，他们的人生常常也过得一团糟。在前信息化时代，由于信息传递的速度受限，全球市场在很大程度上是支离破碎的——这反而给予了创造力更大的生存空间。随着互联网将全球市场越来越整合到一个整体之中，成功的创造有了更强的影响力，但一个创造能够被市场认定为成功的难度同样也变得越来越大。人工智能会将创造力竞争背后的残酷性更加毫无保留地暴露到世人面前。

This is the landscape that emerges when creative competition is coupled with market mechanisms, and it is also a grim reality that our society often conveniently ignores as it continuously mythologizes the quality of creativity. Creation itself is, of course, a beautiful thing, but creation that must stand the test of the market, however, is utterly brutal. The exceptional figures we look up to usually possess immense creativity. But a much larger proportion of talented individuals with equal or even greater creativity failed, for various reasons, to gain acceptance for their creations, and their lives were often a mess.

In the pre-information age, because the speed of information transmission was limited, the global market was largely fragmented. Paradoxically, this gave creativity more room to survive. As the internet increasingly integrates the global market into a single entity, successful creations have gained a more powerful impact, but the difficulty for any single creation to be recognized as successful by the market has also grown ever greater. Artificial intelligence will expose the brutality behind creative competition to the world, even more unreservedly.

从这个意义上，当你听到"人工智能会解放人的创造力"这样的口号时，请不要轻易地将这一描述当作是乌托邦式的未来想象。如果人类当前的经济系统、甚至人类社会的一些底层认知不进行相应的改变，一个"以创造力论英雄"的社会极有可能会是一个比现在更可怕的社会，在这样的社会中极少数人被奉为神明，绝大多数人甚至无法贡献自身的工具性价值。一个真正可以和创造力兼容的未来社会，必然会是一个能够欣赏"无用之用"的社会；现在这个社会距离这一理想设想还太过遥远。

In this sense, when you hear slogans like "AI will unleash human creativity," please do not readily accept this description as a utopian vision of the future. If humanity's current economic system, and even some of the fundamental cognitions of human society, do not undergo corresponding changes, a society that "judges individuals by their creativity" could very well be a more terrifying one than what we have now. In such a society, a tiny minority would be revered as gods, while the vast majority would be unable to even contribute their instrumental value. A future society that is truly compatible with creativity must be one that can appreciate the "usefulness of the useless." Our present society is still far too distant from this ideal.

更何况，如果我的未来预测成立，那么人类的作为"工具性能力"的那部分创造力也必将会被人工智能所超越——这将为人类的工具性价值钉上最后一颗棺材钉。彼时，对于人类而言，唯一重要的创造力就是自我生命表达的创造力，也只有这样的创造力是无法完全外包给人工智能的。人工智能在能够解放人类的这一部分创造力之前，还需要摧毁人类社会几千年来形成的一些根深蒂固的观念。这绝不是仅仅靠替代重复性劳动就能完成的。道阻且长，切莫有太多幻想。

Moreover, if my predictions for the future hold true, then the part of human creativity that functions as an "instrumental ability" will also inevitably be surpassed by AI—this will drive the final nail into the coffin of humanity's instrumental value. By then, for humankind, the only creativity that will truly matter is the creativity of expressing one's own being; and only this kind of creativity cannot be fully outsourced to artificial intelligence. Before AI can liberate this part of human creativity, it must first shatter some of the deep-rooted notions that human society has clung to for millennia. This is a feat that can by no means be accomplished merely by replacing repetitive labor. The path is long and arduous; let us harbor no grand illusions.