顯示具有 AI 標籤的文章。 顯示所有文章
顯示具有 AI 標籤的文章。 顯示所有文章

MAGA+AI並非穩定的良方。最新的科技泡沫只能為川普提供短期保護。

問《金融時報》
美國經濟觀點
最新的科技泡沫只能為川普提供短期保護。

亞當圖茲添加到我的FT

© 伊凡懷特
Maga + AI 並不能保證 x 系統的穩定性(在新視窗中開啟)
MAGA+AI 並不能保證 Facebook 的穩定(在新視窗中開啟)
MAGA+AI 並不能保證 LinkedIn 的穩定(在新視窗中開啟)

分享

節省

專業功能配置
當前進度 100%
亞當圖茲
已發布9小時前
72
列印此頁
作者是《金融時報》的特約編輯,並撰寫 Chartbook 電子報。


在歷史的關鍵時刻,技術變革不僅能帶來經濟成長,還能鞏固或顛覆政治體制。 


第一次世界大戰之後,福特主義不僅是一條生產線,它還提供了大眾富裕的新文化願景。


1970 年代和 1980 年代,微電子和電腦領域的革命標誌著蘇聯集團的終結。


2008年,新一代智慧型手機的問世和社群媒體的興起,大大緩解了金融危機帶來的衝擊。畢竟,如果資本主義為我們帶來了iPhone和Facebook,那它也不全然是壞事。


但這種敘事很容易轉變。傳送帶可以成為疏離的隱喻。數位平台可以成為兩極化和內戰的驅動力。


在我們當下,人工智慧的超大規模應用正蓬勃發展,而唐納德·川普也開始了第二個激進的總統任期。


評論員一致認為,短期來看,這種巧合有助於保護美國總統免受其一些功能失調的政策可能引發的反彈,尤其是來自強大的商業精英的反彈。


企業領導階層和投資者可能會被科技領域的大新聞分散注意力,從而忽略常規公司治理的崩潰。
就美國經濟而言,關稅和移民管制這兩大「弊端」被人工智慧繁榮這一「巨大利好」所抵銷。減稅和股市上漲的雙重利好讓精英階層感到欣慰。


但如果你深入了解,你會發現 Maga 與 AI 之間的平衡所帶來的穩定感其實非常不穩定。


如果川普放任自流的執政是美國社會和政治緊張局勢的產物,那麼「讓美國再次偉大」運動和人工智慧的巧合並不能緩解或減輕這些緊張局勢。

首先,要從普遍存在的國家衰落感說起——這正是美國精英焦慮的根源。拜登政府的因應之策是嚴密控制大型科技公司,限制高科技晶片出口。但這招致了晶片製造商的強烈不滿,反而刺激了中國採取各種巧妙的規避措施。川普政府似乎希望美國能夠透過科技擴散來佔據主導地位。但這往往會加劇衝擊,進一步強化多極格局。然後,

英偉達的黃仁勳告訴你,中國最終還是會贏,那接下來該怎麼辦呢?



在國內,川普政府和拜登政府一樣,公開宣稱的目標是重振美國工業基礎,恢復勞工階級和中產階級的生活穩定。但是,

這又如何與鼓勵矽谷斥資數千億美元用演算法取代大部分白領勞動力的做法相符呢?


這些演算法是一種顛覆性的文化技術。我們最大的希望是,它們能夠實現內生成長理論的願景,並大幅提升研發效率,從而加速以科學為導向的經濟和文化轉型。然而,在一個像川普2.0這樣充斥著蒙昧主義和後真相思想的政府領導下,人工智慧迎來了曙光,這不是一種莫大的諷刺嗎?


無論是JD Vance的新天主教傾向,或是Peter Thiel的怪誕佈道,都無法掩蓋歷史和哲學視野的矛盾之處。


當然,人工智慧或許被高估了,我們不該輕信炒作。但之後會發生什麼事?泡沫會破嗎?

正如吉塔·戈皮納特離開國際貨幣基金組織後指出的那樣,如果美國股市出現類似互聯網泡沫破裂的規模調整,將給美國投資者造成20萬億美元的損失,給世界其他地區造成15萬億美元的損失。這相當於美國GDP的70%和世界其他地區GDP的20%。這足以引發一場嚴重的經濟衰退。我們能想像川普政府在巨額赤字、聯準會政治化和國會僵局的情況下,會如何應對這種情況嗎?


那些仍然抱持著美國有朝一日會恢復理智想法的人,似乎希望出現一次恰到好處的崩盤:

規模要大,但又不能太大——足以消除人工智能的炒作和特朗普執政第一年的迷醉,但又不足以引發比我們正在經歷的更嚴重的憲政危機。


套用九月出版的一本聳人聽聞的人工智慧暢銷書的結尾語,「我真希望我是在誇大其詞」。但如果你希望美國恢復正常,那麼從現在的情況來看,如何才能實現這個目標並不明朗。


版權所有 © 2025 英國金融時報股份有限公司。保留所有權利。
重新利用此內容
(在新視窗中開啟)
評論
跳轉至評論區
美國經濟最新動態
英國就業
英國失業率上升幅度超乎預期,達5%。

房地產行業
隨著美國經濟擠壓老年美國人的生存空間,反向抵押貸款業務略有上升。

美國政治與政策
美國參議院批准結束政府停擺的協議,民主黨面臨黨內強烈反對

川普關稅
在商界人士的推動下,瑞士接近達成協議,將美國對瑞士的關稅降至15%。

化學品
英力士就廉價塑膠進口產品向歐盟提起反傾銷訴訟

美國政府停擺
參議員採取第一步措施結束美國政府停擺

商業機密
最高法院裁決為川普帶來的政治風險
優質的 內容

FT解釋
美國能否打破中國對稀土資源的控制?

請關注本文中的主題
亞當圖茲
添加到我的FT
美國經濟
添加到我的FT
美國政治與政策
添加到我的FT
人工智慧
添加到我的FT
唐納川普
添加到我的FT
評論
評論指南
請保持評論的尊重性。為了方便全球讀者理解,請使用簡潔明了的語言,避免使用可能被誤解為冒犯性的措辭。發表評論即表示您同意遵守我們的社區準則和這些條款及細則。我們鼓勵您舉報不當評論。

發表評論

登入並加入討論

所有評論
73
排序方式

最新

以上皆非'24
4分鐘前
哎呀,振作起來!世界照樣運轉。一切都會好起來的。

推薦

回覆

分享

報告

格莎
9分鐘前
涉及 MAGA 的穩定性只有在前面加上一個「-」號才有可能實現。

推薦

回覆

分享

報告

東京來電
15分鐘前
只要銀行不參與其中(事後我們無法確定銀行是否會參與),世界就能應對“人工智慧崩潰”,就像網路泡沫破滅後實體經濟沒有遭受太大損失一樣。

推薦

回覆

分享

報告

尤里·維齊泰
24分鐘前
文章寫得很好,宏觀視角也很出色。

在微觀層面,MAGA和人工智慧還有其他問題。例如,過去快速識別MAGA狂熱分子很容易:他們無法進行連貫的寫作。 “Borders”就是“Borders”,“Too”和“to”可以互換使用。 「Should of」、「irregardless」和「could care less」等字眼也常被濫用。這是因為作為一個群體,這些人通常不愛讀書。所以,你很快就能分辨出對方是誰,並採取相應的行動。而人工智慧讓MAGA的寫作者聽起來條理清晰、彬彬有禮,賦予他們的表達一種嚴肅的外衣,讓人誤以為對方是一位見多識廣、深思熟慮的人。

這是個實際存在的問題!

推薦

回覆

分享

報告

事實邏輯
1小時前
作者用了多少人工智慧技術來撰寫這篇文章?如果你不理解目前這波人工智慧工具的意義,那麼你完全沒必要使用它們。你會落後於時代,但你並非必須訂閱任何現有的模型。

總之,作者的這句話徹底讓我對他失去了任何認真對待的機會:
關稅和打擊移民的弊端
非法移民就是非法的。如果你想讓它合法化,那就修改法律。否則,就執行你國家現有的邊境法律。

推薦

回覆

分享

報告

胡言亂語
35 MINUTES AGO

 
In reply to FactnLogic
Most Illegal immigrants are actually semi-legal. They pay taxes, have US driving licenses, etc. etc. This is cheap labour.

Recommend
3

Reply

Share

Report

Multipass
22 MINUTES AGO
(Edited)

 
In reply to FactnLogic

Anyway, the author lost any chance of me taking him seriously with this statement:
bads” of tariffs and the crackdown on immigration
The author is simply highlighting the reality that illegal immigrants contribute to the economy and are a major source of cheap labour for US.

As investors, many of us foreign to the US, we benefit from cheap labour pumping our US investments a lot more than we do from immigration enforcement, that is an obvious fact. Reducing the pool of cheap labour will reduce corporate bottom lines, period.

It doesn’t have anything to do with the politics.

Recommend
2

Reply

Share

Report

Tazzyjoolz
1 HOUR AGO
Until AI has proved itself through application in a range of different industries, it will remain exactly what it currently is, a puffball which will disappear when the next best thing comes along.

Recommend
1

Reply

Share

Report

IVA
1 HOUR AGO
AI + CCP is stable.

Recommend
1

Reply

Share

Report

Section8
1 HOUR AGO
The worrying thing about AI is the capitalist greed that is driving it. Such revolutions are unfortunately never driven by altruism. Also, the character profile of some of the MAG7 and AI "giants" CEOs only leads to more concern. Do you really have chosen Zuck or Altman to decide the future of society? They're not the only ones - just the ones most off the rails.
The worrying thing about the current US administration is the current US administration. Between the enrichment focus of the Trump family, the billionaire bros, and the sycophants (hoping to drink at the money or power trough), no-one has the good of the common person at heart.

Recommend
4

Reply

Share

Report

Boulder One
1 HOUR AGO
(Edited)

 
In reply to Section8

You forgot to mention Musk. But I agree. Let’s stop glorifying these guys, who are experimenting on the rest of us and our children.

Recommend
5

Reply

Share

Report

Boulder One
1 HOUR AGO
(Edited)

Here is an unconventional political thought for the AI debate.

AI takes much too much power and the people, who are pushing this on us are greedy villains, who want to line their pockets by destroying our planet.

Since we only use 3% of our brains, let figure out how to use 80% of our brains, become our own versions of AI, and regulate the tech elite.

Maybe we are all getting tired of of the constant manipulation of our species for profit. We are not cattle and this game is getting old..

Recommend
2

Reply

Share

Report

jec
2 HOURS AGO
"Maga + AI is not a recipe for stability"

Blimey, you don't say :)

Recommend
7

Reply

Share

Report

Jonsdottir
3 HOURS AGO
Out of curiosity, I thought I'd have a little look at the value of shares in Duolingo since they declared themselves to be 'AI First' back in April 2025....

https://www.marketwatch.com/investing/stock/duol

...ah.

Recommend
2

Reply

Share

Report

P Martin
3 HOURS AGO
The premise here is that:

1.Trump actually does want to stabilise the working class economic situation
2.The AI fans want to assist humanity rather than chase a dream that reduces the need for humans

Of course the circles can’t be squared…except when you recognise the core of the alliance is between segments ultimately only concerned with the benefits to a white, Neo Christian, ultra wealthy exclusionary sect.

What is good for them is all that matters.

Recommend
6

Reply

Share

Report

George Kaplan
2 HOURS AGO

 
In reply to P Martin
I couldn't agree more. Strangely, political economy seems to be Tooze's blind spot.

Recommend

Reply

Share

Report

Hungry Brain
3 HOURS AGO
"But isn’t there a deep irony in an administration as obscurantist and post-truth as Trump 2.0 presiding over the dawn of artificial general intelligence?"

The lack of ethics and boundaries is actually what a lot of us have been warning against.

Recommend
2

Reply

Share

Report

So, it wasn't that easy
3 HOURS AGO
What is normal?

Recommend

Reply

Share

Report

Jonsdottir
3 HOURS AGO
Nothing to do with this article other than the AI theme, but any thoughts as to why Sarah O'Connor has now been placed behind the Premium Paywall?

Recommend

Reply

Share

Report

Johncy
3 HOURS AGO

 
In reply to Jonsdottir
Terminator?

Recommend

Reply

Share

Report

Jonsdottir
3 HOURS AGO

 
In reply to Johncy
Unlikely, seeing as that is fictional.

Recommend

Reply

Share

Report

chowchow Bubbles
3 HOURS AGO
There was so much that was stated from the past to suggest a technological inflection that made the article interesting to read … I think there may have been a simpler way to link and possibly to tell the story that our governing systems our broken (and this view depends if you are in top 1% or how near it you are or not) and that AI (or at least the current incarnation of AI automation with no regulatory guidance guises in a race with China) will add negatively to the existing problems of our collective society …. Wondering if I understood correctly❓ … 🤔

Recommend
2

Reply

Share

Report

Its Going To Get Worse
48 MINUTES AGO
(Edited)

 
In reply to chowchow Bubbles
Yes you understood perfectly and excellent comment.

Recommend

Reply

Share

Report

Boulder One
4 HOURS AGO
(Edited)

Consider that in the space of two years, AI may become obsolete and be replaced by computers made of human brain tissue or quantum enabling lasers.

Consider that power consumption, based on the new technology may be 10% or even 1% of AI’s power consumption per designated thought process.

What if the new data centers emerge in China, India or are perhaps equally distributed everywhere.

Of course, those who invest in AI will be put out of business. The entire “Big Guy” investment could be replaced by millions of “Small Guy” investments or even a huge “mega-powerful” state actor.

People like to talk about the size of US indebtedness, but what if we invested and additional 35 trillion dollars in something which quickly becomes obsolete or which utterly decimates our social fabric and human personalities in ways, which could never have been imagined.

And, when it comes to “Truth,” my assumption is that we may have already entered a “Post Truth” world, which might be a kind of media-driven LSD Trip, but also filled with a lot of weird ego agendas, which are buried in news articles, which are really advertisements.

We already know that the political world is being pushed forward by a lot of nut-case conspiracy theories and addictive disinformation.

And that brings up the “tip of the iceberg” addictive, drug-like, physiological component of all of this. What does something like Facebook do to emerging adolescent egos and personalities?

It’s going to be a strange ride for sure. Are we not just rats on an endless little dopamine treadmill trips.

Recommend
1

Reply

Share

Report

Its Going To Get Worse
3 HOURS AGO

 
In reply to Boulder One
You may be a rat on an endless little dopamine treadmill trip
but we are Global economists that are Lucid sentient thinking people

We know there going to be financial crises.
Politicians have allowed the wealthy to purloin the money..
By various measures onshore and offshore money laundering schemes.

Recommend
4

Reply

Share

Report

xxzz76
3 HOURS AGO

 
In reply to Boulder One
AI is turbocharged statistical methods.
Some of which have been around for hundreds of years.
It will no more become obsolete than zero - or the calculus.
Its use requires regulation 'tho.

Recommend
2

Reply

Share

Report

Helvetico
4 HOURS AGO
(Edited)
Whining and pining for the fake stability of the old days of Open Borders Neoliberalism is tiresome. Tooze certainly has no answers, other than maybe stick his head in the sand and hope AI goes away. People like him didn't bat an eye when 100's of millions of jobs were exported from the West to the Third World, hollowing out economies and creating an epidemic of deaths of despair. But now he whinges about "elites" (is he not one?) with bursting investment accounts (does he not have them, too?) ignoring the common man. And that common man is somehow a guy who wants even more migrants, legal and illegal, to flood his country, driving up housing prices and pushing down wages, phenomena that once again only benefit the rich.

Tooze is no more an advocate of the common man than Trump, just less entertaining.

Recommend

Reply

Share

Report

Abracadabra
2 HOURS AGO

 
In reply to Helvetico
No big answers in this short piece. But the warning not to be complacent is spot on.

Recommend
2

Reply

Share

Report

DUNJA
5 HOURS AGO
If AI doesn’t generate the returns its founders hope and promise the crash is going to be more than the dotcom crash. It’s going to cause an absolute market meltdown and a financial crisis. The amount of money going into all of these areas is absolutley off the scale.

The reverse is also true if the bet does pay off and AI works in eliminating the need for huge amounts of human labour - there is also going to be an enormous crash and unemployment and sub employment rise exponentially. Either way within next 4-5 years something is going to happen and it’s going to make 08 look small fry in comparison.

Recommend
3

Reply

Share

Report

Swede Wolf
3 HOURS AGO

 
In reply to DUNJA
Of course there will be a spectacular crasch. Where will the revenue stream, justifying current valuations, come from? Not from the people being made redundant I guess. And the affluent have already borrowed to hypercharge their investment in the AI bubble.

Every technological breakthrough is following a similar path. The first phase is filled with optimism and hybris. Then come the crasch to make valuations correspond to reality, and sometimes this overshoot on the downside (and that's when people like Warren Buffet see the opportunity to buy the best companies on the cheap). After that the valuations is determined by real profits and not wishful thinking.

Recommend

Reply

Share

Report

Huginn and Muninn
5 HOURS AGO
(Edited)
" After all, if capitalism gave us the iPhone and Facebook, it could not be all bad. "

True. Generation-wide mental health issues, cohorts of socially inept sociopaths, the industrial scale spread of "learned" ignorance and outright falsehoods galore, and their fuelling of 21st c. fascism are just minor inconveniences.

Recommend
14

Reply

Share

Report

WeAreGrey
4 HOURS AGO
(Edited)

 
In reply to Huginn and Muninn
Capitalism gave us the iPhone and Facebook, enabling better control of the population through mass histeria, division, bread and games. All helping to serve capitalism more, in the form of devaluing labour, and concentrating the remaining capital even more.

But yes, for the rich, it's not all that bad. Just at the cost of everybody else and their souls.

Recommend
1

Reply

Share

Report

p266
60 MINUTES AGO

 
In reply to Huginn and Muninn
It's far worse in most countries, just not so obvious. Speaking from experience.

Recommend

Reply

Share

Report

Reddle
57 MINUTES AGO

 
In reply to Huginn and Muninn
The above sentence looks at it from a more optimistic late 2000s/early-mid 2010s perspective. Back when banks were receiving most of the public's anger and tech companies were fooling people by saying that their platforms would expand our freedoms.

Recommend

Reply

Share

Report

Ragnar_Danneskjold
5 HOURS AGO
I never understand how you can build AI and wipe out all jobs with it in the hopes that this will create growth? there is not growth without a conscientious middle class.

Recommend
9

Reply

Share

Report

Smelly Trumps
5 HOURS AGO

 
In reply to Ragnar_Danneskjold
Stated like a true wage slave - as if employment is the only way to build wealth or have purpose!

Recommend
2

Reply

Share

Report

mailee
5 HOURS AGO

 
In reply to Smelly Trumps
But how will that wealth trickle down to the working and middle classes?

Recommend
1

Reply

Share

Report

Weekender 62
2 HOURS AGO

 
In reply to mailee
Foraging

Recommend

Reply

Share

Report

Ragnar_Danneskjold
5 HOURS AGO

 
In reply to Smelly Trumps
sure that, but do you want to live in a society where wealth creation through a normal job is ruled out from the beginning?

Recommend
3

Reply

Share

Report

xxzz76
3 HOURS AGO

 
In reply to Ragnar_Danneskjold
You can live a very good life without much wealth - it a good, guaranteed and life-long steady income that you need.
As with many Germans, who do not own their homes but have high incomes.

Recommend

Reply

Share

Report

Beobachter, der
4 HOURS AGO

 
In reply to Smelly Trumps
Those "wage slaves" fund nations and governments and their running through their income taxes. Not the companies and the billionaires.
But you obviously punch drunk on learned ignorance of a post-truth world.

Recommend

Reply

Share

Report

Jonsdottir
3 HOURS AGO

 
In reply to Smelly Trumps
Can you build wealth without someone's employment?

Recommend

Reply

Share

Report

xxzz76
3 HOURS AGO

 
In reply to Ragnar_Danneskjold
Growth needs investment.
That does not be necessarily come from the middle classes.
See China...

Recommend

Reply

Share

Report
Show More Replies

Bluenose
5 HOURS AGO
Interesting article. Arguably MAGA (please capitalize) + anything beyond their own destruction is not a good thing.

Too deep a subject to bother with a long text, but does everyone notice that the current ‘crisis’ always the ‘worst crisis’? Frequently it is proclaimed ‘we’re doomed’ and yet…not quite. Arguably Trump’s low IQ, awful personality and slavish following - plus aftershocks when he dies - are a far greater short-term threat than AI.

AI is arguably dangerous, but I was listening this weekend to a group in a hotel lobby holding forth on the subject. An incidence arose of - tragically - a boy committing suicide ‘because AI chat’. Are people aware that many commit suicide because of humans in the dark corners of the internet encouraging them? If you search ‘thinking about suicide’ you will find a lot of homo sapiens with horrific responses. And a lot of good ones trying to discourage.

The steam engine, the telegraph, the internal combustion engine, automation, computers, the internet: so far we have survived. I bloody hate computers, but they are vital.

So maybe have a little hope: it is free and it keeps you warm.

Recommend
1

Reply

Share

Report

Adiaphoros
6 HOURS AGO
Great article. Thanks.

Recommend
4

Reply

Share

Report

wse, Sydney
7 HOURS AGO
MAGA will have a short half life.

Recommend
4

Reply

Share

Report

Cat bot
7 HOURS AGO
“American normality” happened precisely when in the last 300 years? Maybe the 1990s, briefly, but that was more due to global events than anything happening in Washington.

The US is a messy, chaotic, violent, sometimes warmongering, usually innovative experiment. It was never pretty.

Recommend
19

Reply

Share

Report

Ragnar_Danneskjold
6 HOURS AGO

 
In reply to Cat bot
maybe the 50s also

Recommend

Reply

Share

Report

Apples to apples
2 HOURS AGO

 
In reply to Ragnar_Danneskjold
The 50’s were a radical emergence of mobility and youthfulness and federal ambition along with the Korean War 1950-1953, much less staid in the US than is remembered.

Recommend
2

Reply

Share

Report

xxzz76
7 HOURS AGO
AI will yet further dramatically reduce the number of jobs in manufacturing.
What is is proportion of the retail cost of an iPhone that is devoted to assembly (in China/India)?
Less than 5%.
(Production of components requires PhD level designers - but a handful of staff are involved in their automated manufacturing).
As a proportion of tha value added chain,. manufacturing has been declining for many decades.
AI is accelerating this process.
Most of the manufacturing. jobs that Trump is hoping to repatriate to the US are rapidly disappearing - or morphing into far fewer highly skilled jobs that are well beyond the capacities of the Maga mob.
Trump's dream of bringing back an imagined 1950s America is complete fantasy and doomed to failure.
Let's hope the US starts to come back to its senses ..

Recommend
11

Reply

Share

Report

anganaca
6 HOURS AGO

 
In reply to xxzz76
indeed! there seems to be an inherent contradiction between the expectations of Maga people and Trump promises to them and the shredding of jobs AI will create. Income inequality and social polarization from the divergence of skills needed in this new economy will only grow bigger. These two dynamics cannot sit well together.

Recommend
5

Reply

Share

Report

My mate Dave said...
4 HOURS AGO

 
In reply to xxzz76
Exactly. It's the world imagined in Kurt Vonnegut's "Player Piano" playing out (pun somewhat intended).

Recommend
1

Reply

Share

Report

doc martin
7 HOURS AGO
We aren't the first species in the universe to develop AI. WE don't live in a star trek universe. Maybe because AI is the extinction event.

regardless, the evolution of intelligence itself is an extinction level event.

Recommend
2

Reply

Share

Report

SimonJB
7 HOURS AGO
Maga + AI is not a Recipe for Stability

Is stability what is needed right now Adam? China is facing huge domestic problems (lack of growth, youth unemployment, property collapse). Europe is in a proxy war with Russia - for Ukraine this is very much a real war. AI looks like it might be a tranformative technology - and it might affect jobs, especially for the young. That's a huge social change that somehow society will have tk adapt to.

"Cometh the hour, cometh the man"?

Recommend
2

Reply

Share

Report

Jonsdottir
3 HOURS AGO

 
In reply to SimonJB
A quote originating out of a game of cricket...

Recommend

Reply

Share

Report

Bingo69
7 HOURS AGO
A deep one! So unsettling though!

Recommend
1

Reply

Share

Report

Donna C
7 HOURS AGO
Hello, Back in August 2204 I suggested shorting AI stocks. I forgot; buying Put Options would work, too. Seriously, I do want to make two points. FIRST Point. it's my opinion that America's democracy doesn't need any more political messages from political parties or anyone. We need HUMAN messages 'cos AMERICA HAS LOST ITS SOUL. That's the Soul Biden was fighting for. HUMANITY IS the ONLY thing we ALL have in COMMON. Where'd it go ?? It's gotten so beyond nutty that we really need "Back to Basics" classes, across all media, to teach Americans WHAT HUMANITY IS AND HOW TO "SHARE" IT. Call me Crazy. But, You tell me, What Else is Going to Work at this point ?? This is not Rocket Science !!
SECOND Point. AI. Some things that don't bode well. (1) Right off. AI Firms are investing in other AI firms. No portfolio manager with their head screwed on straight would ever invest in a single industry. It's Common knowledge, Fin 101, the whole world knows your RISK skyrockets, if you don't Diversify your Portfolio.
(2) WORSE, each AI Technology Component, be it Chips, Cloud Computing, Algorithms, Data Centers, have serious flaws/points of Failure. They're events with high enough probabilities of occurring that can't be dismissed. Don't even talk about Crypto, BlockChain technology - Decentralized Finance, DeFi. (3) AI Crash ? I'm trying to Time the Market for a crash, but It's a real BEAR. : ) ...couldn't help it, sorry.

Recommend
5

Reply

Share

Report

sven.
7 HOURS AGO
(Edited)
That Trump presides over the dawn of AI is more symptom than irony. The power in automated speech being to create crowds of opinions, but as yet, without any understanding of causal coherence. This is the outcome of automated, high velocity "throw it and see what sticks" being the favoured political persuasion technique.

Then of course there is the challenge presented to rationality by some applications of nanotech as another cause of the philosophical incoherence of various leaders.

Recommend
2

Reply

Share

Report

Paul A. Myers
7 HOURS AGO
But if a return to American normality is what you are hoping for, it is not obvious how you get there from here.
Free and fair elections would work.

But the Supreme Court has been against free and fair elections for a generation. The Supreme Court is also against any Congressionally passed law that constrains presidential authority in any way. Goodbye Posse Comitatus law; hello ICE. The Venezuelans are a preview, not a sideshow.

When constitutionally sanctioned avenues for changing power are blocked, then the political power moves to the streets and revolution. The voter repression on election day next November will be relitigated starting the following day with a counter-revolution because democratic politics will have demonstrated its failure.

The 8 Democratic senators who caved have proved that Washington no longer works and that the reactionary autocracy now fundamentally allocates the resources, wealth, and income of the country. Washington only works for the superrich.

Just how much coup against the constitution do you need before you see the writing on the wall. The future is going to be written with blood in the streets -- just what the constitution was created for to avoid.

Recommend
38

Reply

Share

Report

Semi-retired
4 HOURS AGO
(Edited)

 
In reply to Paul A. Myers
I share your outrage but don't put as much faith in ''blood on the street''. There has been a lot of that in the US over the last 200 years and it didn't make much difference - repression has been stronger. Nor in the UK (I'm currently re-reading UK C18 and C19 history).

Recommend
2

Reply

Share

Report

RealityModeler
4 HOURS AGO

 
In reply to Paul A. Myers
The rise of the robots will ensure that the blood is only from the working and the professional former middle class.

Recommend
2

Reply

Share

Report

CaptainThunder
1 HOUR AGO
(Edited)

 
In reply to RealityModeler
Quite. Speaking of which:
But how does that square with encouraging Silicon Valley to bet hundreds of billions of dollars on replacing a large part of the white collar workforce with algorithms?
There's an old journo adage that if your headline is a question or you leave one hanging off the end of a para, the answer is usually "no". Or in this case, "it doesn't". Decimating the white-collar middle class as "efficiencies" is the unstated but pretty damn obvious aim.

...or at least that's the theory.

In practice, AI is still good at what it was always good at - analysing large datasets and spitting out the best preformatted responses it can. Or blending whatever it can scrape up from source material into slop and hoping the predictive response more or less matches what the client wanted. It's still not good at closing deals (unless they're very simple), looking for glitches, or interacting with the actual humans that want to send credit card or purchase-order details to make you money.

And they're really, really bad at handling "edge cases"/unusual situations, or identifying potential new sources of revenue or problems that might arise later.

So if management are trying to kill jobs with AI, there's a limit to how much they can do. But if your management want to expand what your current workforce can do, taking away the donkey work and letting those users do more valuable tasks requiring human insight, that's something it's much more suited to.

Though that particular message might not go down so well...

Recommend

Reply

Share

Report

Surfnm
7 MINUTES AGO

 
In reply to Paul A. Myers
When constitutionally sanctioned avenues for changing power are blocked, then the political power moves to the streets and revolution.
America is a wild complex place that historically defies easy categorization and attempts to control it. Politics, economics, and civil society are three broad tents where one can watch aspects of different trends and patterns as they play out over time. The current drama in the political tent may influence the events in the other two tents but it does not control them and it is but one tent out of three.

Recommend

Reply

Share

Report

Bbbbb12
8 HOURS AGO
(Edited)
Amazing that the Fed is using up all its tools before

1. AI comes true and destroys white collar jobs

Or

2. AI does not come true and there is a massive wealth destruction event

Oh well I suppose we could have a decade of sub zero interest rates or a Gen Z revolution. Either one would fix the problem

Recommend
10

Reply

Share

Report

xxzz76
3 HOURS AGO

 
In reply to Bbbbb12
AI is destroying white collar call centre jobs on a huge scale already.
(Not publicised, for obvious reasons - insider knowledge).
The UK has an estimated 800,000 call centre jobs.
AI is likely to cull 30% of them.
That means 240,000 fewer jobs.
Nearly a quarter of a million jobs gone - in one sector alone.
Frightening, to say the least.

Recommend
1

Reply

Share

Report

Admiral Chris
8 HOURS AGO
I think you mean Vance's Paleo- Catholicism, not neo

Recommend
26

Reply

Share

Report

AI Generated Comment
2 HOURS AGO

 
In reply to Admiral Chris
Exactly

Recommend

Reply

Share

Report

David Hume
1 HOUR AGO

 
In reply to Admiral Chris
Postliberalism

Recommend

Reply

Share

Report

Gottgruss
8 HOURS AGO
Interesting read, thought provoking, a troubling start to my day… thanks Mr Tooze.

Recommend
23

Reply

Share

Report

Munk
8 HOURS AGO

 
In reply to Gottgruss
Combines nicely with today’s article “Investor angst over Big Tech’s AI spending spills into bond market”

Recommend
6

Reply

Share

Report

Fair capitalist
9 HOURS AGO
Great article, and a nice, albeit unsettling, summary of the times.

Recommend
62

Reply

Share

Report

Gobsmacked old man
8 HOURS AGO

 
In reply to Fair capitalist
Yes!

Recommend
1

Reply

Share

Report

CaptainThunder
54 MINUTES AGO
(Edited)

 
In reply to Fair capitalist
Came here to say that, and just add one thing...
How do we imagine that the Trump administration, with a gaping deficit, a politicised Fed and a deadlocked Congress, would react to that?
...there's something that needs expanding out of that.

The senior members of the Trump admin are hand-picked for loyalty and obedience. Independence, smarts, and quick thinking are not in their wheelhouse. During Trump 1.0, there were quite a number of clowns in the circus but some actual talents too. Jim Mattis and Mark Milley thankfully kept any fat fingers off the big boom buttons: Alex Azar and Rob Redfield alongside the much-maligned Fauci at Health and CDC: and while I loathe Mnuchin for multiple different reasons, he was at least reasonably sane and kept Navarro on a shorter leash.

This time around... there are. no. adults. in. the. room.
None.*

None of this lot have the chops to deal with even a small crisis, let alone work together to do so. Hegseth is doing lethal "firework shows" while insulting combat-tested officers: FDR's brainworm is rapidly dismantling DoH: Navarro and charlatans of every stripe have a hall pass to the Oval Office…
…and Bessent has let an a number of colossal bubbles inflate (not just AI, credit markets too), with no visible attempt at restraint or enforcing due diligence.

If that bubble bursts (still an if, but the vibe shift has been rather noticeable lately), it'll only directly affect a more limited sector than 2000 or 2008. But with so much market concentration and now so much debt, a lot more investors and pensioners will get hit than in 2000.

*Technically Stephen Miller may be. I'm less familiar with his species, and I'd have to check in the hive he's been building under the old Rose Garden to see what their gestation lifecycle's like. Still doesn't behave like an adult human, though, and if I was his Overlord I'd have him fed to the hatchery for bringing my invasion force into disrepute. Call that a disguise? Frell me dead, c'mon, man...

Recommend
4

Reply

Share

Report

Top of comments

Top of page
Useful links
Support
View Site Tips
Help Centre
聯絡我們
關於我們
無障礙
我的FT之旅
職業生涯
供應商
法律與隱私
條款及細則
隱私權政策
Cookie政策
管理 Cookie
版權
奴隸制聲明及政策
服務
安全分享新聞線索
個人訂閱
專業訂閱
重新出版
高階主管職位搜尋
在《金融時報》刊登廣告
關注 FT 在 X 上的報導
FT頻道
FT學校
工具
資料夾
FT App
FT數位版
FT 編輯
警報中心
商學院排名
訂閱管理器
新聞推播
新聞簡報
貨幣轉換器
社區及活動
FT 現場活動
FT論壇
FT領導力學院
FT集團的更多內容
市場數據至少延遲15分鐘。 © 2025年金融時報股份有限公司。 FT和「金融時報」是金融時報有限公司的商標。 《金融時報》及其新聞報導受《金融時報編輯守則》的
自我監管機制約束。
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%

如果中國贏得人工智慧競賽會怎樣?美國應力爭勝利,但也要做好屈居第二的準備



跳至內容
電腦電路板上的半導體晶片,2022 年 2 月
電腦電路板上的半導體晶片,2022 年 2 月弗洛倫斯·羅/路透社
亞當‧西格爾 (ADAM SEGAL)是美國外交關係協會艾拉‧A‧利普曼新興科技與國家安全主席,也是數位與網路空間政策計畫主任。 2023年4月至2024年7月,他曾擔任美國國務院網路空間與數位政策局高級顧問。
塞巴斯蒂安·埃爾鮑姆是維吉尼亞大學電腦科學教授、外交關係委員會駐地技術專家、電腦機械協會和電氣電子工程師協會會員。
科技高層、國家安全分析師和美國官員似乎都一致認為,美國必須在人工智慧競爭中戰勝中國。 2024 年 10 月,拜登政府國家安全顧問傑克·沙利文警告稱,如果美國不“更快、更全面地部署人工智慧以加強國家安全”,美國就有可能“浪費來之不易的領先地位”。川普政府在其首批行政命令之一中宣布,其目標是「維持和增強美國在全球人工智慧領域的主導地位」。華盛頓為爭奪霸權採取了雙管齊下的策略:透過限制關鍵技術零件的出口來限制中國,並加速國內基礎人工智慧模型的創新。為了實現後一個目標,兩屆政府都對產業領袖實施了相對寬鬆的監管,有針對性地投資半導體和能源基礎設施,並鼓勵聯邦政府,尤其是國防和情報機構採用人工智慧,用於調查食源性疾病爆發和檢測金融詐欺等多種用途。
迄今為止,這些政策使美國公司在市場份額和模型性能方面保持領先於中國同行。但華盛頓不能也不應該期望其領先地位永遠持續下去。 DeepSeek、阿里雲、百度和騰訊等中國人工智慧公司近期的突破表明,美國與中國在尖端人工智慧領域的差距正在縮小,而美國在人工智慧領域的霸主地位遠未穩固。即使華盛頓繼續爭奪主導地位,也需要為未來美國在人工智慧競爭中敗給中國——或者至少,中國人工智慧模型在全球範圍內同樣受歡迎——的可能性做好準備。 
然而,為美國可能屈居第二做好準備,並不意味著華盛頓注定要重蹈5G競爭的覆轍。在5G競爭中,美國一直難以提供價格合理的創新產品,而中國則一路領先。相反,透過推廣更豐富的框架,使其能夠充分體現人工智慧對新興市場最具吸引力的特徵,使開發人員更容易在模型之間遷移,構建允許用戶比較競爭人工智慧模型輸出的系統,與模型構建者安全地共享美國數據,並與合作夥伴和盟友攜手合作,華盛頓可以確保,即使其在爭奪全球霸權的道路上未能成功,也能安全地從人工智慧革命中受益。 

吹鉛

直到2024年夏天,美國似乎掌握著人工智慧主導地位的致勝法寶。快速的模型創新源自於美國生態系統,該生態系統由學術研究、出版物和人才、私營部門資本和開發以及極寬鬆的監管監督組成。自2022年ChatGPT首次席捲全球以來,OpenAI的GPT和Google的Gemini等美國基礎模型已展現出顯著的進步。在過去兩年中,人工智慧工具減少了「幻覺」(這個術語用於描述人工智慧產生虛假或矛盾資訊的現象),獲得了處理和生成圖像和聲音的能力,掌握了更複雜的任務,並展現出增強的推理能力。包括Anthropic、xAi和Meta在內的更多美國科技公司迅速採取行動,開發更強大的模型,這些模型能夠勝任諸如理解語音和圖像、編程以及快速解決複雜科學問題等基準任務。隨著美國模型展現出其速度和精湛技藝,其全球市場份額和國際用戶迅速增長。 
到2023年底,領先的美國模型在響應準確率方面已領先中國模型兩位數個百分點。但中國憑藉一系列明智的政府舉措,迅速縮小了差距。這些措施包括制定《下一代人工智慧發展規劃》,強調人工智慧教育和人才培訓,加大科學研究投入,加強政府與科技產業的密切合作,以及在資料中心、能源傳輸和半導體製造領域投入大量資金。這些努力幫助美國和中國人工智慧工具在最熱門基準測試中的表現差距在2024年底前縮小到個位數。在過去六個月中,DeepSeek和Qwen的表現已追平美國模型,這引發了人們對美國曾經遙遙領先的地位已不復存在的擔憂。 
同時,中國在將人工智慧融入高科技製造業方面處於領先地位。例如,最初以智慧型手機製造商起家的小米,在其北京工廠使用了700多個人工智慧引導機器人,平均每76秒就能生產一輛新的電動車。人工智慧在中國城市廣泛應用於交通管理、監控和執法,各省市政府正在試行建立人工智慧創新區,以促進政府治理、醫療衛生和教育領域的新應用。 
此外,事實證明,美國出口管製作為阻止中國採購先進晶片的壁壘,其有效性遠低於許多政策制定者和產業分析師的預期。北京利用空殼公司和囤積晶片來規避管制,並加速了其國內晶片研發項目。中國企業也率先採用軟體開發技術,最大限度地利用現有硬件,從而優化訓練和推理時間並降低整體成本。無論這些進步是否最終使中國在人工智慧領域處於領先地位,美國在人工智慧領域絕對佔據主導地位的時代顯然已經結束。 

充分利用第二

美國人工智慧公司在建立基礎人工智慧模型的許多方面仍可能保持全球領先地位。 OpenAI 與軟銀和甲骨文合作,於今年 1 月宣布了一項名為「星際之門」的 5,000 億美元人工智慧基礎設施投資項目。亞馬遜、Meta、微軟和谷歌也都持續向新創公司、人工智慧實驗室和人才投入數十億美元。亞馬遜 AWS、微軟 Azure 和谷歌雲端佔據了全球雲端市場 60% 以上的份額,而雲端市場是建置和部署模式的關鍵資源,而中國頂級供應商阿里巴巴的市場份額僅為 4%。然而,隨著人工智慧前沿探索的難度加大、美國在全球人才競爭中的下滑以及聯邦研究經費削減對創新的抑制,過去三年推動美國企業的技術創新步伐可能落後於中國,甚至難以為繼。華盛頓應該努力避免這種情況的發生。白宮科技辦公室已開始製定新的國家人工智慧行動計劃,預計將於 7​​ 月發布。 
儘管如此,政策制定者也應該為一個競爭激烈的人工智慧生態系統共存的世界做好規劃。值得慶幸的是,即使美國未能直接贏得創新競爭,華盛頓也能找到其他策略,確保其從人工智慧進步中獲益。 
首先,美國應該尋找新的方式向全球市場展示其模型的優點。美國國家標準與技術研究院可透過人工智慧安全研究所及其產業合作夥伴,推廣基礎人工智慧模型的新評估架構。標準基準主要關注語言理解、推理和對話能力等原始能力,而忽略了其他指標,例如模型的透明度、可追溯性和可訪問性、運營成本以及模型權重(用於微調模型以做出準確預測的“旋鈕”)的修改難易程度。納入這些指標的新評估框架可以用來吸引新的市場和用戶,即使美國企業的模型在標準基準上無法持續勝過中國模型,也能保持其競爭力。
隨著更多模型的出現,用戶將希望避免在美國或中國現有產品之間做出選擇(並因此被鎖定)。在即將到來的人工智慧消費者選擇時代,最大限度地降低模型之間的遷移成本將成為一個有吸引力的賣點。美國人工智慧產業可以透過降低採購價格並減少在模型之間遷移所需的軟體修改、硬體升級和人員培訓量來降低向其模型過渡的成本和複雜性。華盛頓可以領導國際標準化組織(ISO)的努力,對基礎模型的應用程式介面(允許不同軟體程式進行通訊和交換資料的協定)進行標準化,這可以降低模型之間遷移的成本,並減少對任何一個國家人工智慧產品的依賴。如果中國模型確實佔據領先地位,那麼給予全球用戶一定程度的信心,讓他們相信自己可以利用多種模型的優勢,如果美國再次成為第一,他們也可以重新轉向美國模型,這將是一個明智的選擇。
政策制定者也應該為一個相互競爭的人工智慧生態系統共存的世界做好規劃
隨著中國(以及最終其他)模式日益強大並滲透全球市場,華盛頓不能僅僅強調審查和間諜活動的風險,就指望美國和外國公司拒絕採用這些模式。相反,美國公司必須建立基於基礎模型的系統和應用程序,但要降低依賴任何特定模型的風險。在應用程式和基礎模型之間添加另一個軟體層(稱為中間抽象層),可以將下游系統與基礎模型隔離,使其更加獨立和更具彈性。如果基礎模型發生更改,對運行在其上的應用程式或其他模型的效能產生負面影響,或者出現了新的、更好的基礎模型,那麼建立下游應用程式的公司可以迅速轉向其他基礎模型。
美國用戶和企業採用中國的基礎模型確實會帶來實際風險,包括容易受到不正確、扭曲、剝削性甚至破壞性輸出的影響;敏感資料暴露給潛在對手;以及可能導致整個經濟部門癱瘓的服務中斷威脅。在基於外國基礎模型進行建置時,美國企業需要確保其下游應用能夠抵禦此類威脅。美國企業應該建立裁決系統,該系統能夠比較功能較弱但值得信賴的模型與功能較強但不受信賴的模型的響應,確定不受信賴模型的響應是否令人滿意,警告用戶潛在風險,並防止使用不正確的響應。此類系統會權衡相互競爭的回應,從而增加開發和維護成本,並降低迴應時間,因此美國政策制定者應優先將裁決系統整合到受不可靠外國技術引入影響最大的下游應用中,例如醫療診斷、詐欺識別和交通管制。
最後,美國應該規範美國開發商和公司與外國模型建構者共享的資料類型,但不應默認全面禁止資料傳輸。出於隱私和國家安全原因,華盛頓對與中國共享美國數據持謹慎態度,這可以理解,但在某些情況下,使用美國數據對中國模型進行微調所帶來的經濟或社會效益可能大於風險。例如,如果中國出現了一種人工智慧工具,其醫療診斷和治療建議的準確性遠高於美國同類工具,那麼美國醫院就應該使用中國模型,即使冒著共享個人患者資訊的風險,甚至可能共享更多患者群體的數據,這些數據可用於根據美國人群對模型進行微調。在這些情況下,美國公司可以透過資料匿名化、資料脫敏(以虛假或加密資料取代敏感資料)以及差異隱私(允許在群體間共享資料同時限制個人資訊共享的數學框架)來降低隱私和安全風險。
華盛頓希望標準化新的評估指標,並制定與盟友和朋友分享數據的指南。它還需要向缺乏專業知識的合作夥伴提供技術和資金援助,幫助他們在不同模型之間遷移,並建立和應用用於裁決競爭模型的系統。 

迷失方向

華盛頓面臨全球人工智慧格局的演變,絕對主導地位已不再穩固。美國政策制定者不能僅僅依靠沙文主義的號召來贏得人工智慧競賽,而忽視美國早期領先地位不會永遠保持的可能性。當然,華盛頓仍然應該努力保持領先。一個更負責任、更現實的策略是推行有利於美國繁榮發展的政策,同時為美國未能實現絕對主導地位做好準備。 
否則,華盛頓將面臨最糟糕的結果:一個更勝一籌的競爭對手,其經濟和軍事實力在人工智慧的推動下不斷增強,而美國國內的人工智慧產業卻無力跟上,因為在必要時無法借鑒中國的模式。屈居第二並非美國人工智慧的喪鐘,但拒絕適應競爭才是。 

您正在閱讀一篇免費文章

訂閱《外交事務》即可取得無限制存取權限。
  • 免費閱讀新文章和一個多世紀的檔案
  • 每年有六期印刷版和線上版,還有音頻文章
  • 解鎖《外交事務》應用程序,隨時隨地閱讀
已經是訂閱者?登入

ForeignAffairs.com 上的 Cookie

本網站使用 Cookie 來提升您的使用體驗。您可以使用 Cookie 管理頁面選擇停用某些 Cookie。
哎呀,出了點問題。請查看瀏覽器的開發者控制台以了解更多詳細資訊。

韓國AI人才戰略:邁向全球AI三強的策略與願景

韓國AI人才戰略:邁向全球AI三強的策略與願景

作者:劉羽芹/工研院產科國際所政策組助理研究員


在全球AI競爭加劇背景下,AI技術已成各國科技競爭核心。韓國面對美中等強國壓力,為在未來科技領域佔有一席之地,對AI人才培育展現高度重視。隨著傳統產業持續數位化轉型,AI技術不僅成為推動產業升級與經濟成長的核心動力,也為應對少子化與高齡化所引發的勞動力短缺提供解方。因此,加速培育高階AI人才已成為韓國科技戰略中的首要任務。

韓國當前以「2030年躋身全球AI三強」為目標,積極推動AI人才政策。然而,韓國現階段AI人才儲備嚴重不足。韓國科學技術情報通信部在政策報導中指出,2023年韓國AI人才數僅約5.1萬名,距離韓國政府2030年20萬名AI人才的需求目標尚有顯著差距。為此,韓國正積極推動人才政策,致力於縮小與全球領先國家的差距。然而,韓國AI技術與美國和中國相比,亦存在一定差距。EPOCH AI在2024年7月的統計顯示,美國以擁有64個生成式AI模型排名第一,其次是中國的42個模型,韓國則以11個模型排名第三。此外,韓國目前尚未出現AI獨角獸企業,仍需更多資源支持初創企業的快速發展。為彌補AI人才缺口,韓國政府計劃於2030年前培育超過250家提供AI自主製造解決方案的企業,積極推動AI人才政策,以促進AI技術的普及及人才培育。

 

AI人才政策:打造AI技術強國

韓國為實現2030年躋身全球AI三強的目標,推行一系列旨在提升國際競爭力的AI人才政策,以在全球AI競爭中取得優勢地位。其中,以下幾項關鍵政策將成為提升其AI全球競爭力的推動力:

  • 「國家AI戰略政策方向」:由韓國科學技術情報通信部主導之國家層級AI發展戰略,涵蓋產業、教育、人才等方面,推動韓國成為全球AI領先國家。計劃於2024至2027年間吸引65兆韓元(約新臺幣1兆4,950億元)民間投資,GPU算力於2030年提升至2EF以上,推動國產AI半導體商業化,並聚焦AI安全研究與規範應用推動,目標為實現70%的產業應用及95%的公部門AI導入率。為應對日益嚴重的深度偽造犯罪與網路安全威脅,韓國於2024年11月成立「AI安全研究所」,並於同年12月通過「AI基本法」,以全面保障AI技術的安全與規範化應用。為促進AI人才與產業的有效銜接,建立完善的AI人才資料庫IRIS,追蹤國家AI研發計畫人員就業流動狀況,助力政府制定有效AI人才政策,促進產學結合。此政策預計在2026年創造310兆韓元(約新臺幣7兆1,300億元)經濟效益。

  • 「基於數據的科技人才政策深化戰略」:韓國科學技術情報通信部以數據分析為基礎,精準制定科技人才政策,為《國家AI戰略政策方向》提供人才支持。通過大數據分析掌握科技人才需求,追蹤理工碩博士發展,並建立涵蓋12項國家戰略技術領域的人才地圖,進一步強化產學合作,緩解人才供需失衡,提升研發投資效率。

  • 「科學·數學·資訊·跨學科教育綜合計畫」:韓國教育部推進AI數位教科書的應用,結合探究式教學模式,致力於全面培養學生的科學素養、數學邏輯思維與計算能力。該計畫以提供個人化學習方案與專任教師專業培訓為基礎,融入STEAM教育理念及跨學科專案,強化學生解決創新問題的能力。此外,計畫涵蓋科學教育推廣活動,旨在整合多方資源,全面提升數位教育體系的國際競爭力(詳見圖一)。

韓國科學·數學·資訊·跨學科教育綜合計畫融入STEAM教育理念,強化學生解決創新問題的能力。詳細內容可參考韓國科學技術情報通信部公告

圖一、科學·數學·資訊·跨學科教育綜合計畫方向

資料來源:韓國科學技術情報通信部,2024/12;作者重製圖,2024/12

  • 「AI自主製造戰略0」:由韓國產業通商資源部主導,專注於製造業的AI應用。2024年首年度投入超過1,000億韓元(約新臺幣23億元),推動200項AI自主製造專案,吸引1兆韓元(約新臺幣230億元)的民間投資,培育13,000名專業人才與250家專門企業。此計畫目標至2030年,將AI自主製造普及率從目前的9%提升至30%以上,並使製造生產力提高20%以上。此計畫旨在推廣AI自主製造技術,提升製造業整體效率,進一步鞏固韓國在全球AI製造領域的領導地位。

 

多管齊下推進AI布局

韓國政府推行AI政策的同時,亦採取多項措施促進人才培育與技術應用。例如,鼓勵產學合作,支持大學與研究機構設立AI研究基地,如「國家AI研究基地」,專注於突破性AI研究,包括超越神經網路規模化法則及機器人基礎模型研究等。此外,韓國政府計劃開發適用於中小學的個人化AI數位教科書,預計2025年推出數學與資訊版本,2027年擴展至科學領域,並自2026年起逐步應用於課後融合教育課程的開發。

 

韓國模式與我國借鑑

韓國在AI人才發展積極目標下的舉措,對我國有借鑑價值。首先,可考慮建立國家級AI研究基地,集中資源推動突破性AI研究。同時,加強AI在中小學教育的滲透,例如開發並應用AI數位教科書,逐步將其納入中小學課綱,以普及AI教育,培養學生的AI思維與實踐能力。為促進AI人才與產業的有效銜接,可建立完善的AI人才資料庫,系統化追蹤與分析相關數據,為人才政策制定與大學課程設置提供依據,進一步推動學產協同發展。通過借鑑韓國經驗並結合我國國情,可逐步完善我國AI人才發展戰略,實現人才培育與產業需求的協同,為未來在國際科技競爭中奠定穩健的基礎。

 

參考資料:

  1. 韓國教育部,《以AI與數位技術支援學生及教師的個人化成長——推動跨領域人才培育》,2024年12月20日。

  2. 韓國科學技術情報通信部,《邁向全球AI三大強國的新里程碑——韓國國家AI研究中心成立》,2024年10月28日。

  3. 韓國科學技術情報通信部,《邁向全球AI三大強國的藍圖——韓國AI創新戰略發布》,2024年9月26日。

  4. 韓國科學技術情報通信部,《實現全球頂尖科學技術人才大國——「基於數據的科學技術人才政策深化戰略」發布》,2024年6月4日。

  5. 韓國產業通商資源部,《AI時代來臨——勾勒韓國產業的未來方向》,2024年5月7日。

  6. 韓國科學技術情報通信部,《政府與民間攜手探索邁向全球AI三大強國之路——「AI戰略高層協議會」成立》,2024年4月3日。

核電復興又帶來一個老問題:如何處理核廢料


穆罕默德·佐杜阿 (Mohammed Zoeduah) 在新澤西州卡姆登的 Holtec International 製造廠焊接一個可容納乏核燃料數十年的罐體

核電復興又帶來一個老問題:如何處理核廢料


A Nuclear-Power Revival Brings Back an Old Problem: What to Do With the Waste

每年都有數噸放射性廢棄物暫時存放

Tons of radioactive refuse goes into temporary storage every year
穆罕默德·佐杜阿 (Mohammed Zoeduah) 在新澤西州卡姆登的 Holtec International 製造廠焊接一個可容納乏核燃料數十年的罐體
經過
埃里克·尼勒
 |照片和影片由喬卡羅塔為《華爾街日報》拍攝
2025年3月5日上午5:30(美國東部時間)
穆罕默德·佐杜阿 (MOHAMMED ZOEDUAH) 在新澤西州卡姆登的 HOLTEC INTERNATIONAL 製造廠焊接一個可容納乏核燃料數十年的罐體
核電復興(部分原因是由耗電龐大的人工智慧資料中心推動)再次引發了一個棘手的問題:如何處理遺留下來的放射性廢棄物。
目前,已有超過 9 萬公噸的乏核燃料被存放在 39 個州的場所。根據太平洋西北國家實驗室 2024 年的報告,其中包括 73 座商業核電廠和 30 多所大學和政府設施。
自 1960 年代美國核工業開始起飛以來,核廢料就一直堆積在用於臨時儲存的乏燃料池和乾式貯存罐中,而美國能源部未能按照法律要求永久處置這些核廢料,導致納稅人花費數百億美元來補償這些公用事業公司。
將核廢料儲存在遠離反應爐地點的新措施也遭遇了阻力。週三,最高法院將審理一宗案件,決定私人公司是否可以在德州和新墨西哥州的設施中暫時儲存乏燃料。
州政府官員和其他人士認為,核管管理委員會在授予這些設施許可證時超越了其權限。預計該案要到今年稍後才會做出判決。
同時,核反應器持續提供美國近20%的電力,每年產生約2,000公噸廢棄物。隨著越來越多的工廠投入使用以滿足資料中心、工業廠房、家庭和電動車的需求,廢物堆積量將進一步增加。  
美國三十年來首批新反應器於去年在喬治亞州竣工。目前正計劃重新開放密西根州、愛荷華州、賓州南卡羅來納州已關閉的反應爐。
川普總統已承諾加快審批仍在研發的 新型小型模組化核反應器的許可證。
核能之所以具有吸引力,是因為它幾乎不產生溫室氣體,幾乎可以提供不間斷的電力。但根據對科學家和工程師的採訪,它產生的放射性廢物應該永久儲存在地下深處。
喬治亞州沃格特爾核電廠的兩座反應爐中的第二座將於 2024 年開始運作。這是 30 年來美國首次建造的新反應器。
曾於2012年至2014年擔任美國核管理委員會主席、現任不列顛哥倫比亞大學公共政策與全球事務學院院長的艾莉森·麥克法蘭表示:“如果沒有解決廢物問題,就繼續建造新反應堆是不負責任的。”
其他核子國家正在加快建設核廢料處置庫。
芬蘭正在對一個地下核電廠進行最終測試,該核電廠位於該國西南部一片森林下方 1500 英尺處,預計將於 2026 年開放。該國擁有五座反應堆,約佔全國發電量的三分之一。
法國70%的電力來自核能,約96%的廢棄物被回收製成新燃料,其餘的則儲存在諾曼第的一個中央冷卻池中。法國官員預計將於 2027 年開始在法國東北部建造一個永久性地下儲存庫,並於 2035 年開放。 
11 月,加拿大官員選定了安大略省西北部基岩中核廢料的永久存放地點。
但美國的努力在幾十年前就停滯了,根據政府問責局2021 年的一份報告,當時在內華達州尤卡山建造永久性地下儲存設施的 150 億美元項目因該州民選官員的反對而被迫停止。
根據對核子專家、政策制定者和民選官員的採訪,尤卡山核事故給建設永久性處置場的努力蒙上了一層陰影。 
美國能源部負責乏燃料和高放射性廢物處置的副助理部長保羅·默里 (Paul Murray) 在該機構 11 月 26 日主辦的網絡研討會上表示:“不存在技術問題,也不存在工程問題。” “我們真正想做的是建立公眾信任和政治意願,以便真正採取行動。我們已經完成了所有需要做的研究。”
在永久性儲存設施建成之前,核退役公司 Orano USA 與德州廢棄物處理公司廢棄物控制專家的合資企業 Interim Storage Partners 提議將多達 40,000 公噸的乏核燃料暫時儲存在德州一處目前存放低放射性廢棄物的設施的乾式儲存桶中。
另外,核子工程服務公司 Holtec International 提議在新墨西哥州建立一個地下乾式儲存設施,該設施距離德克薩斯州的設施約 40 英里,可儲存 8,680 公噸乏核燃料。
2023 年,聯邦巡迴法院站在德州官員和其他反對者一邊,裁定 NRC 超越了其權限,授予了臨時儲存合作夥伴在其場地儲存核廢料 40 年的許可證,並有可能延長許可證期限。同一家法院也於 2024 年阻止了 Holtec 在新墨西哥州設立臨時儲存站的 NRC 許可證。
最高法院將於週三審理 NRC 對德州裁決的上訴。最高法院的裁決也將決定新墨西哥州工廠是否能夠繼續前進。
即使建造了這些臨時儲存場,乏燃料中的一些元素仍會在數千年內保持放射性,而聯邦法律要求永久儲存場必須持續 10,000 年。這意味著仍然需要一個永久性的地質儲存庫——某個深而乾燥且遠離活躍地震斷層的地方。
根據美國能源資訊署的數據,一個典型的核反應器擁有 121 至 193 個燃料組件,其中裝有 12 英尺長的鈾燃料棒。每個組件有179至264根燃料棒。五到七年後,鈾開始降解,每隔 18 個月,大約三分之一的燃料組件會從反應器轉移到乏燃料池。 
典型的游泳池深度約為 40 英尺,面積可達 40 平方英尺或更大。牆壁厚四英尺至八英尺,由鋼筋混凝土製成,內襯不銹鋼。廢物儲存在至少 20 英尺深的水下。
勞倫普林斯 (LAUREN PRINCE) 和以賽亞愛德華茲 (ISAIAH EDWARDS) 在新澤西州卡姆登的 HOLTEC INTERNATIONAL 製造廠組裝乏燃料容器。隨著核能需求的增加,該公司的乏燃料容器銷售量也不斷上升。
隨著水池被注滿,乏燃料組件被轉移到乾式貯存桶中。 
霍爾特克國際公司 (Holtec International) 在費城市中心特拉華河對岸的一家大型工廠內製造的桶由不銹鋼罐組成,這些不銹鋼罐嵌套在較厚的碳鋼外殼內,中間澆注了兩英尺厚的混凝土。整個裝置重 360,000 磅,最多可容納 89 個乏燃料組件。相較之下,乏燃料池可容納多達數千個乏燃料組件。
根據法律規定,能源部必須接收和儲存乏核燃料,但由於永久性處置設施從未建成,政府因未能處置廢料而每年向公用事業公司支付 6 億至 8 億美元的賠償金。 
自 1998 年以來,美國能源部已向公用事業公司支付了 111 億美元的賠償金。根據美國能源部監察長辦公室 2024 年的審計結果,該機構未來在廢棄物處理方面的負債預計將高達 445 億美元。
「我們將需要更多的核能,」德州共和黨眾議員奧古斯特·普弗魯格 (August Pfluger) 表示,他共同發起了一項法案,旨在成立一個新機構來尋找永久性處置場。 “但我們確實存在浪費問題。
致函 Eric Niiler,信箱:eric.niiler@wsj.com

版權所有 ©2025 道瓊斯公司。保留所有權利。 87990cbe856818d5eddac44c7b1cdeb8

加入對話即表示您接受我們的社群規則和條款。若要查看規則、條款和常見問題解答 點這裡。如有疑問,請發送至moderator@wsj.com。

新聞文章和新聞專欄的討論必須圍繞著故事主題。在觀點內容中,對話可以深入探討其他主題。關於這篇文章的討論將在發表四天後結束。

第一個發表評論...
似乎還沒有人就此話題發表自己的看法留下評論,以便您的聲音能夠被首先聽到。

選擇汪精衛中華帝國會像奧匈帝國鄂圖曼土耳其帝國一樣戰敗解體

選擇汪精衛 中華帝國會像奧匈帝國鄂圖曼土耳其帝國一樣戰敗解體 因為站錯了隊伍 北洋軍閥頭腦比汪精衛清楚 所以一戰才能拿回山東 孫文拿德國錢,他是反對參加一戰 選擇蔣介石, 中國將淪為共產主義國家 因為蔣介石鬥不過史達林 蔣介石即使打贏毛澤東 中國一樣會解體 中國是靠偽裝民族主義的...