AI的問題，不是發展還是停止

The Real Question of AI Is Not Whether to

Develop or Stop

——從一位年輕人的激情發言，看生命、智能與組織信託

—From a Young Person’s Passionate Speech to Life, Intelligence, and Organizational Trust

錢 宏Archer Hong QIAN

2026年9月18日晨於溫哥華

2026 年 9 月 15 日倫敦 Canva“AI Vision”會上，Pull The Plug 一位年輕女性站上座位發言，隨後被安保帶離。她講的核心不是抽象的“AI 會不會滅絕人類”，而是更落地的一層：地方社區最先承受數據中心對電、水、土地和住房的擠壓，卻幾乎沒有表決權。

這段倫敦AI前景會議上的視頻傳開，不是因為她發明了新問題，而是她把已經存在多年的張力壓進了一個具體房間：台上在談“願景”，台下在談“誰先付賬”。面對台上的AI產業人士，激動地談到AI給普通人和地方社區帶來的影響。

她年輕、勇敢，聲音富有感染力，對普通人的生活懷有強烈關切。我很欣賞這樣的年輕人。一個社會如果還有年輕人願意站起來，對正在改變人類生活的巨大力量公開表達自己的憂慮，這本身就是一種可貴的責任心、愛心和公共精神。

視頻下面的評論很快形成了截然不同的聲音。

有人讚揚她是“清醒者”，有人認為她是在反對AI；有人把問題歸結為資本利用AI剝削普通人；也有人反駁：AI研發的巨額投入和失敗風險主要由資本承擔，投資者當然應該擁有相應的經營決策權。還有人認為AI已經不可阻擋，人類真正需要做的是適應，而不是恐懼。

這些評論讓我想到：我們可能首先需要改變討論AI的方式。

AI真正的問題，並不是贊成還是反對、發展還是停止，也不是讓公眾通過一次民主表決來決定AI究竟應該向前還是向後。

因為AI首先是一件必須通過一定組織形式才能完成的事情。

那麼，真正值得追問的就是：

做AI的組織，究竟是不是一個忠於信託的組織？

一、究竟是誰在推動AI？

表面上看，推動AI的是科技巨頭、企業家、資本家和工程師。

但再往後看一步，就沒有這麼簡單。

今天許多重要AI企業背後連接着複雜的資本市場。股東、基金、養老金、機構投資者以及無數普通投資者，通過不同渠道把自己的資源投入這些企業。

一個普通人買進一家科技公司的股票，當然不能解釋成他對這家公司每一項AI決策都投了一張贊成票。但是，他願意把自己的錢交給這個組織，本身就是一種經濟意義上的選擇和託付：

我把自己的資源交給你，希望你運用這些資源創造價值。

因此，所謂“資本推動AI”，並不準確，因為其背後實際上連接着千千萬萬普通人的儲蓄、投資和未來預期。

政府也是如此。

當AI已經關繫到產業競爭、生產效率、醫療、教育、國家安全以及一個國家未來的科技能力時，政府不可能簡單地干預說：“因為存在風險，所以停止發展。”

政府本身也是受託組織。

它既要防範風險，也必須考慮：怎樣使自己的人民不因為技術變化而失去未來的發展機會？怎樣避免自己的國家在新的生產力競爭中陷入被動？

而普通公眾更不是AI的局外人。

我們使用互聯網，貢獻數據，購買產品，投資企業，繳納稅收，同時又生活在AI正在改變的社區、就業市場、教育體系和社會環境之中。

所以，AI背後真正站着的，並不只是幾個億萬富翁。

資本、技術人員、政府、投資者、消費者、社區與普通人的生活，實際上已經連接在同一條巨大的組織鏈條之中。

二、AI爭議的深處，是組織信託

如果我們把AI問題簡單變成“資本家與人民”“科技巨頭與普通人”“技術進步與人類安全”的二選一，很容易再次掉進熟悉的二元對立。

資本承擔投資風險，當然應該獲得相應的產權和創造回報；但是，投入資本並不意味着同時取得處置全部社會外部性的無限權利。

公眾受到AI影響，當然應該擁有與自己生命和生活相關的權利；但是，公眾的焦慮也不能自動成為判斷複雜技術未來的充分依據。

政府應該監管；但監管者本身同樣需要接受信託檢驗。

科技企業需要創新；但創新者也必須回答：你所使用的資本、能源、水、土地、數據以及組織權力，最終創造了什麼，又向誰承兌？

所以問題逐漸清楚了：

誰受託發展AI？

受誰之託？

使用什麼資源？

創造什麼價值？

由誰承擔外部成本？

最終又向誰承兌？

這些已經不只是技術監管問題。

這是組織信託問題。

也因此，我們真正需要建立的，不是簡單的“支持AI聯盟”或者“反對AI聯盟”，而是讓生命、人工智能與組織信託——LIFE—AI—TRUST——發生交互契合。

三、AI的第一個瓶頸，不只是耗電

這位年輕女士的發言以及由此引起的討論，還提醒了我們過去容易被低估的問題。

AI消耗的不只是電。

它需要數據中心、芯片、土地、水資源、礦產、網絡，需要大量工程師和其他專業人員，需要巨額資本，也需要越來越龐大的能源和基礎設施體系。

因此，我過去概括的AI第一個瓶頸——能耗與能效不對稱——還可以進一步展開為：

資源—能耗投入與生命能效產出的不對稱。

我們不能只問：

模型擴大了多少？

算力增加了多少？

運行速度提高了多少？

AI產業又創造了多少GDP？

我們還應該問：

它為生命降低了多少成本？

賦予普通人多少新的創造能力？

改善了多少健康？

增加了多少人與組織之間的信任？

減少了多少衝突？

這也正是為什麼我一直主張，在GDP之外還需要看GDE：

C——降本，E——賦能，H——健康，T——信任，P——和平。

如果越來越多的電、水、土地、礦產、資本和人的創造力，只是為了把某一種局部能力不斷推向極致，那麼即使技術指標一次又一次刷新紀錄，我們仍然應該問一句：

生命究竟得到了什麼？

四、AI最容易被忽略的問題：片面智能的極大發展

這可能是今天AI討論中一個更加根本、卻很少被真正說清楚的問題。

我們習慣說“人工智能越來越接近人的智能”，甚至不斷追問：

AI什麼時候超過人類？

可是，今天所謂Artificial Intelligence，並不是把一個完整的“人”複製進機器。

它實際上是把人的某些能夠數據化、形式化、計算化、模式化的智能能力抽取出來，再通過數據、算法、算力和神經網絡，把這些能力放大到過去難以想象的程度。

所以，今天發生的首先不是“完整人的智能”被整體人工化。

更準確地說，是：

人的某些智能被人工化，並獲得了極其強大的片面發展。

這一點，與現代社會的專業分工和現代知識體系的學科分割有一種驚人的相似。

經濟學研究經濟，醫學研究身體，政治學研究權力，工程學研究效率，計算機科學研究計算。每一個格子都可以越來越精細，每一種專業能力都可以越來越強大。

可是，人並不生活在任何一個格子裡面。

生命是整體發生的。

AI也面臨同樣的問題。

當某些局部智能越來越強，而生命的整體性沒有同步進入時，片面發展的力量越強，它與真實生命世界之間的不匹配反而可能越明顯。

於是，第一個瓶頸與更深層的問題連接起來了：

為什麼AI需要越來越巨大的資源投入？我們究竟是在用越來越多的生命資源，把什麼樣的智能推向極致？

五、第二個瓶頸：AI面對的，本來就不是完整世界

AI的信息從哪裡來？

來自人類已經生產、記錄和保存的數據。

但是，這些數據在進入AI以前，就已經經過人的觀察、認知、分類、選擇和價值判斷。

世界上無數沒有被記錄的經驗，無法被量化的感受，沒有進入數據庫的生命活動，在AI開始訓練之前，就已經不在它的“世界”之中了。

然後，AI再對這些已經經過選擇的信息進行訓練、計算、組合和生成。

因此，從信源—信道—信果來看，AI面對的從來就不是未經選擇的完整世界。

這就是我所說的第二個瓶頸：

系統思維的局限。

局部的、經過篩選的信息，即使經過再強大的計算，也不會僅僅因為算力增加，就自動變成完整的生命世界。

於是，我們不能只問AI的答案聰明不聰明。

還應該繼續問：

它從哪裡得到這個“真”？

它怎樣理解“善”？

它怎樣感受“美”？

它又憑什麼生成“慧”與“信”？

當我們一路從真、善、美追問到慧與信，問題已經越出了單純的Artificial Intelligence。

它開始進入Mind。

六、第三個瓶頸：Intelligence不等於Mind

今天圍繞AI最大的焦慮之一，是它會不會有一天反過來控制人類。

這樣的風險當然可以研究，也應該研究。

但是，從“AI某些智能能力超過人類”，直接跳到“AI將成為一個完整的、獨立於人的新統治主體”，中間其實存在一個巨大的認識鴻溝。

因為：

數據+算法+算力+神經網絡，並不因此就等於Mind，更不等於Amorsophia。

Intelligence可以被工程化，可以被訓練，可以被複製，可以在某些任務上不斷放大，甚至遠遠超過任何一個人的能力。

可是Mind究竟是什麼？

人的意識、生命體驗、愛、價值感、意義感，以及主體與主體之間的交互與生成究竟怎樣發生？

這些問題，人類自己都還遠遠沒有完全弄清楚。

因此，我們尤其需要警惕一種認知錯位：

AI真正值得警惕的，不只是人工智能越來越強，而是人類把越來越強的局部智能，誤認為越來越完整的Mind。

一旦發生這種錯位，就可能同時產生兩種看起來完全相反的情緒：

一種是無限崇拜AI。

另一種是無限恐懼AI。

一個認為AI無所不能，一個擔心AI終將成為人類無法控制的新主體。

方向完全相反，卻可能來自同一個認識前提：

把局部智能能力的極大發展，等同於完整Mind的生成。

在這個意義上，對不起，我不能不說，連AI領軍人物及諾獎得主辛頓先生也未能免俗。

七、她並未提出新問題，但放大了老問題

現在再回到那位年輕女士。

我仍然欣賞她。

但我們需要準確地說，她並未提出一個新的問題，這些擔憂早已存在。但她以年輕人的激情、鮮明的表達和對普通人及社區生活的關切，把一個已經持續多年的老問題重新放大了。從早期關於AI失控風險的警告，到要求暫停更強大模型訓練的公開呼籲，再到近年來圍繞AI安全、對齊與監管不斷出現的爭論，這些問題早已由AI科學家、企業家和研究人員自己提了出來。今天真正值得思考的，已經不是再把“AI危險不危險”“應該發展還是停止”變成一次大眾的贊成與反對，而是怎樣把已經被反覆提出、不斷放大的問題重新定義準確，並進一步尋找可以解決它的路徑。

從生成式AI迅速發展的早期開始，AI企業家、科學家和研究人員內部就不斷有人公開討論模型風險、安全、對齊、就業衝擊以及監管。此後，來自AI企業內部的安全爭論、公開分歧乃至研究人員辭職，又不斷把這種憂慮推到社會面前。

但是，這裡恰恰出現了一個更需要我們警惕的問題：

問題被放大，並不等於問題被準確地定義。

如果一個複雜問題被不斷放大，最後卻被壓縮成：

你支持AI還是反對AI？

你站科技企業還是站普通人？

發展還是停止？

資本還是人民？

自由創新還是政府監管？

那麼，一個原本需要深入認識和解決的問題，就可能變成要求所有人立即選擇立場的問題。

社會情緒越強烈，問題本身反而可能越來越模糊。

這也是為什麼，我認為今天真正需要做的，已經不是繼續增加“AI很危險”或者“AI不可阻擋”的音量。

我們需要重新定義問題。

AI並不是一個通過一次民主表決就可以決定前進或者停止的東西，也不是一個只靠幾個科學家、企業家或者政府官員就可以完成的工程。

它已經成為由資本、科研、工程、能源、水資源、土地、數據、政府、企業、投資者、消費者和社區共同參與的巨大組織活動。

因此，真正的問題不是：誰對，誰錯？

甚至首先也不是：誰贊成，誰反對？

而是：怎樣解決？

八、不必停下來，但必須升格

到了這裡，我們才可能重新理解這些年來不斷出現的AI焦慮。

真正需要解決的，是兩組相互連接的問題。

第一組，是AI自身的三大瓶頸：

資源—能耗投入與生命能效產出的不對稱；

系統思維的局限；

Artificial Intelligence與Mind之間尚未跨越的鴻溝。

第二組，則是發展AI的組織信託問題：

誰受託？

受誰之託？

怎樣使用資源？

怎樣創造價值？

怎樣承擔風險和外部成本？

怎樣讓創造出來的價值重新賦能生命？

最終向誰承兌？

這樣一來，“監管還是不監管”也不再是問題的終點。

監管當然需要。

但是誰監管？監管什麼？依據什麼價值尺度？監管者向誰負責？如果監管本身成為新的高成本組織，誰又來檢驗監管者？

於是問題再次回到：組織信託。

這也就是為什麼，我越來越認為，面對AI，我們既沒有必要因為恐懼而停下來，也不能因為競爭而閉着眼睛向前衝。

不必停下來，但必須升格。

從 AI for All 走向 AI for Life。

從單純追求Artificial Intelligence的能力擴張，走向尊重Mind、服務生命的AM（Artificial Mind & Amorsophia MindsField / Network）。

從“誰控制AI”的權力爭奪，走向 LIFE—AI—TRUST 的交互契合共生。

從數據、算法、算力和神經網絡構成的技術系統，進一步追問生命、Mind、Amorsophia以及人與人之間怎樣建立可信的受託關係。

這不是要消滅Artificial Intelligence。

恰恰相反，是要讓它找到自己在生命世界中更加準確的位置。

結語：把被放大的焦慮，變成可以解決的問題

所以，我感謝那個年輕人。

不是因為她告訴了我們AI究竟應該發展還是停止，也不是因為她給出了AI問題的答案。

她讓我們再次看見了一個已經存在、卻仍然沒有得到很好解決的社會焦慮。她把問題放大了！

接下來真正重要的工作，是不要繼續放大情緒，而要把被放大的問題重新放回準確的問題域。

AI當然需要發展，也當然需要面對風險。

但比“支持還是反對”更重要的是認識它；比“樂觀還是悲觀”更重要的是解決它；比簡單地增加監管更重要的，是建立能夠承擔責任、接受驗證並持續向生命承兌的組織信託。

因為真正的問題最終不是：

AI會不會成為人類的敵人？

而是：

我們正在用什麼認識、什麼價值、什麼資源和什麼樣的組織形式，把AI帶進人類生活？

如果只是把人的某些智能不斷人工化、無限放大，我們得到的可能只是越來越強大的Artificial Intelligence。

如果讓智能重新回到生命，讓技術接受組織信託，讓 LIFE—AI—TRUST 交互契合，我們才有可能從 AI for All 進一步走向 AI for Life。

AI真正需要跨越的，不只是下一代模型（https:// amorsophia.com，ai-am.org）。

人類真正需要完成的，也不是下一場關於AI的贊成與反對。

需要發生的，是從片面智能走向生命整體，從AI焦慮走向組織信託，從Artificial Intelligence走向：

Double AM ——Artificial Mind & Amorsophia MindsField / Network.

AI老問題被提出 → 女孩把它放大 → 大眾情緒進一步放大 → 發現原來的問題域本身有局限 → 重新定義AI三大瓶頸 → 從監管進入組織信託 → 從AI for All進入AI for Life → 將AI升格為AM。

Elevation is not an upgrade.

這才是《將AI升格為AM》真正要解決的問題。

The Real Question of AI Is Not Whether to Develop or Stop

—From a Young Person’s Passionate Speech to Life, Intelligence, and Organizational Trust

Archer Hong QIAN
Vancouver, Morning of September 18, 2026

At Canva’s “AI Vision” event in London on September 15, 2026, a young woman from Pull The Plug stood on a chair to speak and was subsequently escorted out by security. The core of her message was not the abstract question of whether “AI will destroy humanity,” but something much closer to everyday life: local communities are often the first to bear the pressure that data centres place on electricity, water, land, and housing, yet they have little say in the decisions.

The video spread widely not because she had invented a new problem, but because she compressed years of accumulated tension into one physical room: onstage, people were talking about “vision”; offstage, someone was asking, “Who pays first?”

Facing AI industry participants onstage, she spoke passionately about the effects of AI on ordinary people and local communities.

She was young and courageous. Her voice was compelling, and she showed strong concern for ordinary people’s lives. I appreciate young people like her. A society in which young people are still willing to stand up and publicly express their concerns about forces transforming human life possesses something precious in itself: responsibility, 愛, and a public spirit.

The comments beneath the video quickly divided into sharply different camps.

Some praised her as someone who was “awake.” Some interpreted her speech as opposition to AI. Some reduced the issue to capital using AI to exploit ordinary people. Others responded that the enormous investments and risks of failure in AI development are borne largely by capital, and that investors should therefore possess corresponding decision-making rights. Still others argued that AI is already unstoppable and that humanity should adapt rather than fear it.

These reactions made me think that perhaps the first thing we need to change is the way we discuss AI.

The real question of AI is not whether we support or oppose it, whether it should develop or stop, nor whether the public should decide through a single democratic vote whether AI should move forward or backward.

AI, first of all, is something that can only be accomplished through particular forms of organization.

The question truly worth asking, therefore, is:

Are the organizations developing AI organizations that remain faithful to trust?

I. Who Is Actually Driving AI?

On the surface, AI is being driven by technology giants, entrepreneurs, capitalists, and engineers.

Look one step further, however, and the picture becomes much more complicated.

Behind many of today’s major AI companies lies a complex capital market. Shareholders, funds, pension funds, institutional investors, and countless ordinary investors channel their resources into these companies through different mechanisms.

When an ordinary person buys shares in a technology company, this certainly cannot be interpreted as casting a vote in favour of every AI decision made by that company. Yet the willingness to entrust one’s money to an organization is itself an economic choice and a form of entrustment:

I entrust my resources to you in the expectation that you will use them to create value.

It is therefore not entirely accurate to say simply that “capital is driving AI.” Behind capital are the savings, investments, and future expectations of millions of ordinary people.

The same applies to governments.

Once AI affects industrial competition, productivity, healthcare, education, national security, and a country’s future technological capabilities, a government cannot simply intervene by saying: “There are risks, therefore development must stop.”

Government itself is an entrusted organization.

It must guard against risks, but it must also ask: How can it prevent its people from losing future opportunities because of technological change? How can it keep the country from falling into a disadvantaged position in the competition surrounding new productive capabilities?

Nor is the general public standing outside AI.

We use the internet, contribute data, purchase products, invest in companies, and pay taxes. At the same time, we live within the communities, labour markets, educational systems, and social environments that AI is transforming.

So the forces behind AI are not merely a handful of billionaires.

Capital, technical professionals, governments, investors, consumers, communities, and ordinary people’s lives have already become connected through one enormous organizational chain.

II. Beneath the AI Controversy Lies Organizational Trust

If we reduce the AI question to binary oppositions—“capitalists versus the people,” “technology giants versus ordinary citizens,” or “technological progress versus human safety”—we easily fall back into familiar dualistic thinking.

Capital bears investment risk and should therefore receive corresponding property rights and returns from creation. But investing capital does not confer an unlimited right to dispose of all social externalities.

The public is affected by AI and therefore should possess rights concerning its own life and livelihood. But public anxiety cannot automatically become sufficient grounds for judging the future of complex technologies.

Government should regulate. Yet regulators themselves must also be tested against trust.

Technology companies need to innovate. But innovators must also answer: What do the capital, energy, water, land, data, and organizational power you use ultimately create—and to whom are those values fulfilled?

The questions gradually become clearer:

Who is entrusted to develop AI?
Entrusted by whom?
What resources are being used?
What value is being created?
Who bears the external costs?
And to whom is that value ultimately fulfilled?

These are no longer merely questions of technological regulation.

They are questions of organizational trust.

What we truly need, therefore, is neither a simple “pro-AI alliance” nor an “anti-AI alliance,” but an interactive congruence among Life, Artificial Intelligence, and Organizational Trust—LIFE—AI—TRUST.

III. AI’s First Bottleneck Is Not Merely Electricity Consumption

The young woman’s speech and the discussion surrounding it also remind us of something that has often been underestimated.

AI consumes far more than electricity.

It requires data centres, chips, land, water, minerals, networks, large numbers of engineers and other professionals, enormous amounts of capital, and an increasingly vast system of energy and infrastructure.

The first bottleneck of AI that I have previously identified—the asymmetry between energy consumption and energy efficiency—can therefore be further expressed as:

the asymmetry between resource-energy inputs and life-efficiency outputs.

We cannot ask only:

How much larger has the model become?

How much computing power has been added?

How much faster does it run?

How much GDP has the AI industry created?

We must also ask:

How much has it reduced the cost of life?

How much new creative capacity has it given ordinary people?

How much health has it improved?

How much trust between people and organizations has it generated?

How much conflict has it reduced?

This is precisely why I have long argued that, beyond GDP, we also need to consider GDE:

C—Cost Reduction; E—Empowerment; H—Health; T—Trust; P—Peace.

If increasing amounts of electricity, water, land, minerals, capital, and human creativity are devoted merely to pushing one localized capability toward its extreme, then even if technical benchmarks repeatedly break records, we must still ask:

What has life actually gained?

IV. AI’s Most Easily Overlooked Problem: The Extreme Development of Partial Intelligence

This may be an even more fundamental problem in today’s AI debate, yet one that is rarely articulated clearly.

We routinely say that “artificial intelligence is getting closer to human intelligence,” and repeatedly ask:

When will AI surpass humanity?

But today’s Artificial Intelligence does not place a complete human being inside a machine.

What it actually does is extract certain human capabilities that can be datafied, formalized, computed, and patterned, and then amplify them through data, algorithms, computing power, and neural networks to levels previously unimaginable.

What is happening, therefore, is not primarily the artificial reproduction of complete human intelligence.

More precisely:

Certain human intelligences have been artificialized and developed to extraordinary extremes in partial form.

There is a striking similarity here to modern specialization and the disciplinary fragmentation of knowledge.

Economics studies the economy. Medicine studies the body. Political science studies power. Engineering studies efficiency. Computer science studies computation. Every box can become increasingly refined, and every specialized capability increasingly powerful.

But human beings do not live inside any one of those boxes.

Life happens as a whole.

AI faces the same problem.

When particular forms of intelligence grow ever more powerful while the wholeness of life does not enter the process correspondingly, the stronger these partial capabilities become, the more visible their mismatch with the real world of life may become.

The first bottleneck therefore connects with a deeper question:

Why does AI require ever greater resource inputs? What kind of intelligence are we using ever more life resources to push toward its extreme?

V. The Second Bottleneck: AI Never Faces the Complete World

Where does AI obtain its information?

From data already produced, recorded, and preserved by human beings.

But before these data ever reach AI, they have already passed through human observation, cognition, classification, selection, and value judgment.

Countless experiences that were never recorded, feelings that cannot be quantified, and life activities that never entered databases have already disappeared from AI’s “world” before training even begins.

AI then trains, calculates, combines, and generates from information that has already undergone selection.

Therefore, viewed through the relationship among information source—information channel—information outcome, AI never confronts an unselected and complete world.

This is what I call the second bottleneck:

the limitations of systemic thinking.

Partial and filtered information does not automatically become the complete world of life simply because more computing power is applied to it.

So we cannot ask only whether AI’s answers are intelligent.

We must continue asking:

Where does its “truth” come from?

How does it understand “goodness”?

How does it experience “beauty”?

And on what basis could it generate “wisdom” and “trust”?

Once we move from truth, goodness, and beauty toward wisdom and trust, we have already moved beyond Artificial Intelligence alone.

We have entered the domain of Mind.

VI. The Third Bottleneck: Intelligence Is Not Mind

One of the greatest anxieties surrounding AI today is whether it might eventually turn around and control humanity.

Such risks can and should be studied.

But there is an enormous conceptual gap between saying that “some AI capabilities exceed those of human beings” and saying that “AI will become a complete new ruling subject independent of humanity.”

Because:

Data + Algorithms + Computing Power + Neural Networks ≠ Mind, much less Amorsophia.

Intelligence can be engineered. It can be trained and copied. It can be amplified continuously in particular tasks and can far exceed the capability of any single human being.

But what exactly is Mind?

How do human consciousness, lived experience, 愛, values, meaning, and the interaction and generation between subjects actually occur?

Human beings themselves are still far from fully understanding these questions.

We therefore need to be particularly alert to a cognitive misalignment:

What deserves vigilance is not merely that Artificial Intelligence is becoming stronger, but that human beings may mistake increasingly powerful partial intelligence for an increasingly complete Mind.

Once this misalignment occurs, it can simultaneously produce two apparently opposite emotions:

One is unlimited worship of AI.

The other is unlimited fear of AI.

One assumes that AI can do everything; the other fears that AI will inevitably become a new subject beyond human control.

The directions are completely opposite, yet both may arise from the same cognitive premise:

equating the extreme development of partial intelligent capability with the generation of a complete Mind.

In this sense, I am sorry to say that even AI pioneer and Nobel laureate Geoffrey Hinton has not entirely escaped this assumption.

VII. She Raised No New Problem, but She Amplified an Old One

Now let us return to the young woman.

I still appreciate her.

But we should describe what happened accurately. She did not raise a new problem. These concerns had existed for years. What she did—with youthful passion, vivid expression, and concern for ordinary people and community life—was to amplify an old problem that had already persisted for a long time.

From early warnings about the risk of AI losing control, to public calls for pauses in the training of more powerful models, and then to continuing debates over AI safety, alignment, and regulation, these questions had already been raised by AI scientists, entrepreneurs, and researchers themselves.

What deserves serious thought today is no longer how to turn “Is AI dangerous?” or “Should AI develop or stop?” into a mass vote of approval or opposition. It is how to redefine repeatedly raised and amplified problems accurately, and then find pathways capable of addressing them.

Since the rapid rise of generative AI, entrepreneurs, scientists, and researchers within the AI industry have repeatedly discussed model risk, safety, alignment, employment disruption, and regulation. Internal safety disputes, public disagreements, and even resignations by researchers have subsequently pushed these anxieties further into public view.

But here lies something that requires even greater vigilance:

Amplifying a problem does not mean defining it accurately.

If a complex problem is amplified again and again, only to be compressed into:

Do you support AI or oppose it?

Do you stand with technology companies or ordinary people?

Development or cessation?

Capital or the people?

Free innovation or government regulation?

then a problem that originally required deeper understanding and resolution may become a demand that everyone immediately choose a side.

The stronger the social emotion becomes, the less clear the problem itself may become.

This is why I believe what we need today is no longer simply to increase the volume of either “AI is dangerous” or “AI is unstoppable.”

We need to redefine the problem.

AI is not something whose advance or cessation can be decided by a single democratic vote. Nor is it an undertaking that can be accomplished solely by a few scientists, entrepreneurs, or government officials.

It has become an enormous organizational activity involving capital, research, engineering, energy, water, land, data, governments, businesses, investors, consumers, and communities.

The real question, therefore, is not:

Who is right and who is wrong?

Nor, first of all:

Who supports it and who opposes it?

It is:

How do we solve the problem?

VIII. No Need to Stop—but We Must Elevate

Only now can we begin to understand the AI anxieties that have repeatedly surfaced over recent years.

What truly needs to be addressed consists of two interconnected sets of problems.

The first is AI’s three intrinsic bottlenecks:

the asymmetry between resource-energy inputs and life-efficiency outputs;
the limitations of systemic thinking;
the still-unbridged gap between Artificial Intelligence and Mind.

The second is the question of organizational trust in developing AI:

Who is entrusted?
Entrusted by whom?
How are resources used?
How is value created?
How are risks and external costs borne?
How can created value empower life again?
And to whom is it ultimately fulfilled?

Seen this way, “regulation or no regulation” is no longer the endpoint of the question.

Regulation is of course necessary.

But who regulates? What is regulated? According to what scale of value? To whom are regulators accountable? And if regulation itself becomes another high-cost organization, who examines the regulators?

Once again, the question returns to:

organizational trust.

This is why I increasingly believe that, in confronting AI, we neither need to stop because of fear nor should we rush blindly forward because of competition.

No need to stop—but we must elevate.

From AI for All toward AI for Life.

From the mere expansion of Artificial Intelligence capabilities toward AM—respecting Mind and serving life:

Double AM—Artificial Mind & Amorsophia MindsField / Network.

From struggles over “who controls AI” toward the interactive congruence and symbiosis of:

LIFE—AI—TRUST.

From technical systems composed of data, algorithms, computing power, and neural networks toward deeper questions of life, Mind, Amorsophia, and how trustworthy entrusted relationships can be established among human beings.

This does not mean eliminating Artificial Intelligence.

Quite the contrary.

It means helping AI find its more accurate place within the world of life.

Conclusion: Turn Amplified Anxiety into a Solvable Problem

So I thank that young person.

Not because she told us whether AI should develop or stop, and not because she gave us the answer to the AI problem.

She enabled us to see once again a social anxiety that has existed for years but has still not been adequately resolved.

She amplified the problem.

What matters next is not to amplify the emotion further, but to return the amplified problem to its proper problem domain.

AI certainly needs to develop, and it certainly needs to confront risk.

But understanding it matters more than “supporting or opposing” it. Solving it matters more than being “optimistic or pessimistic.” And more important than simply adding regulation is building organizational trust capable of bearing responsibility, accepting verification, and continuously fulfilling its trust toward life.

Because ultimately the real question is not:

Will AI become humanity’s enemy?

It is:

With what understanding, what values, what resources, and what forms of organization are we bringing AI into human life?

If all we do is continuously artificialize and infinitely amplify certain human intelligences, what we may obtain is simply ever more powerful Artificial Intelligence.

If we bring intelligence back to life, place technology within organizational trust, and enable LIFE—AI—TRUST to interact and achieve congruence, then we may be able to move from AI for All toward AI for Life.

What AI truly needs to cross is not merely the threshold to the next generation of models.

What humanity needs to accomplish is not another contest between approval and opposition to AI.

What needs to happen is a movement:

from partial intelligence toward the wholeness of life;
from AI anxiety toward organizational trust;
from Artificial Intelligence toward:

Double AM—Artificial Mind & Amorsophia MindsField / Network.

The old AI problem is raised → the young woman amplifies it → public emotion amplifies it further → we discover the limitations of the original problem domain → redefine AI’s three bottlenecks → move from regulation toward organizational trust → move from AI for All toward AI for Life → elevate AI to AM.

Elevation is not an upgrade.

This is precisely the problem that Elevating AI to AM seeks to address.