权力与算力同桌：G7 的 AI 午餐会，和一场没有结论的较量

六月十七号，在法国小镇依云，

On June 17, in the small French town of Évian,

G7 峰会的最后一天，

on the final day of the G7 summit,

办了一场特别的工作午餐会。

a special working lunch was held.

我们先交代一下背景。

Let's first lay out the background.

G7，也就是"七国集团"，

G7 — the 'Group of Seven' —

是由美国、英国、法国、德国、

is made up of the US, UK, France, Germany,

日本、意大利、加拿大，

Japan, Italy, and Canada,

这七个发达国家组成的。

these seven developed countries.

每年，这些国家的领导人，

Every year, these countries' leaders

都会聚到一起，

gather together

讨论全球最重要的议题。

to discuss the world's most important issues.

而今年，最受关注的议题，

And this year, the most-watched issue,

毫无悬念，是人工智能，也就是 AI。

with no suspense, is artificial intelligence — AI.

这场午餐会真正特别的地方在于它的座位。

What truly made this lunch special was its seating.

餐桌的一边，

On one side of the table

是掌握政治权力的各国元首，

were the heads of state who hold political power,

包括美国总统特朗普、

including US President Trump

法国总统马克龙等人。

and French President Macron, among others.

另一边，

On the other side

则坐着一群掌握着技术权力的人——

sat a group of people who hold technological power —

也就是当今全球最顶尖的 AI 公司的掌门人。

the leaders of today's top global AI companies.

据报道，出席的科技领袖将近十几位，

According to reports, nearly a dozen tech leaders attended,

其中包括 OpenAI 的奥特曼、

including OpenAI's Altman,

谷歌 DeepMind 的哈萨比斯、

Google DeepMind's Hassabis,

Anthropic 的阿莫迪，

Anthropic's Amodei,

还有 Meta、Mistral、Cohere、

plus executives from Meta, Mistral, Cohere,

Salesforce 等公司的高管。

Salesforce, and others.

可以说，全球 AI 的半壁江山，

You could say half of global AI

都坐在了这一张餐桌旁。

was seated at this one table.

这场午餐会的主题，

The theme of this lunch,

官方的说法是：

in the official wording, was:

"确保人工智能安全、快速、

'Ensuring artificial intelligence is deployed safely, rapidly,

而有效地落地。

and effectively.'

说白了，桌上的人，

Put plainly, the people at the table

都在试图回答同一个问题：

were all trying to answer the same question:

当 AI 强大到这个程度，

when AI is this powerful,

我们该如何驾驭它？

how do we steer it?

讨论的话题很广。

The topics discussed were wide.

从最前沿的 AI 风险，

From the most frontier AI risks,

到支撑 AI 的基础设施，

to the infrastructure that supports AI,

再到一个国家的"技术主权"，

to a country's 'tech sovereignty',

甚至包括如何在网络上，保护儿童。

and even how to protect children online.

但是，整场午餐会最核心的张力，

But the core tension of the whole lunch

其实集中在一个问题上：

was concentrated on one question:

究竟，由谁来为 AI 制定规则？

who, in the end, sets the rules for AI?

在这一点上，

On this point,

几位美国科技公司的掌门人，

several American tech-company leaders

给出了相当一致的方向。

gave a fairly consistent direction.

Anthropic 的阿莫迪，

Anthropic's Amodei,

和谷歌 DeepMind 的哈萨比斯，

and Google DeepMind's Hassabis,

都呼吁建立一个，

both called for building

由美国主导的 AI 联盟，

a US-led AI coalition —

由美国牵头，联合一批国家，

with America taking the lead, uniting a group of countries,

共同来制定 AI 的标准和规则。

to jointly set AI's standards and rules.

OpenAI 的奥特曼，

OpenAI's Altman

则进一步提议，

went a step further, proposing

建立一个国际性的论坛，

an international forum,

为各国的 AI，

to give each country's AI

提供统一、公正的能力与风险测试，

a unified, impartial test of capability and risk,

并成为各国合作的平台。

and to be a platform for cooperation among nations.

这些提议听起来很有建设性。

These proposals sound constructive.

但是，如果你站在欧洲的角度，

But if you stand in Europe's shoes,

味道就有点不一样了。

the flavor is a little different.

近年来，以法国为代表的欧洲国家，

In recent years, European countries led by France

对美国在 AI 领域的绝对主导，

have grown more and more wary

越来越警惕。

of America's absolute dominance in AI.

他们反复强调一个概念：

They repeatedly stress one concept:

"技术主权"。

'tech sovereignty'.

言下之意是：

What they imply is:

欧洲不愿意，

Europe is unwilling

只是被动地接受美国的技术、

to passively accept America's technology,

美国的标准、美国定下的游戏规则。

America's standards, the game rules America sets.

法国方面就明确表示，

France stated clearly

欧洲要的是有能力去竞争，

that what Europe wants is the ability to compete,

而不仅仅是去合规、去跟随。

not merely to comply and follow.

所以你看，

So you see,

这场看似其乐融融的午餐，

this seemingly warm and friendly lunch,

桌面之下，

beneath the table,

其实是一场关于

was a quiet contest over

未来话语权的暗自角力。

future say-so.

一边，是技术的创造者，

On one side, the creators of the technology,

希望由自己、由自己的国家，

hoping that they, and their own country,

来主导规则的制定。

lead the making of the rules.

另一边，是其他的国家，

On the other, other countries,

不愿意把自己的未来，

unwilling to hand their future

完全交到别人手里。

entirely into someone else's hands.

那么，这场午餐会，

So, this lunch —

最终谈出什么结果了吗？

did it ultimately produce a result?

答案是：几乎没有。

The answer is: almost none.

会后，没有发布任何具体的政策，

Afterward, no specific policy was released,

也没有形成任何有约束力的协议。

and no binding agreement was formed.

所有人只是坐下来，

Everyone just sat down

把各自的立场和担忧，

and put their positions and worries

摆到了台面上，如此而已。

on the table — that was all.

但是，我想说，

But, I'd say,

不要小看这样一场"没有结论"的午餐。

don't underestimate such a 'resultless' lunch.

因为它本身，

Because it itself

就传递了一个清晰的信号：

sends a clear signal:

人工智能，

artificial intelligence

已经不再只是科技公司的事情了。

is no longer just a matter for tech companies.

它的重要性，

Its importance

已经上升到，

has risen to the point

需要全球最有权力的政治领袖，

where it needs the world's most powerful political leaders,

和最有能力的技术领袖，

and its most capable technological leaders,

面对面、坐下来，

to sit down face to face

郑重其事地讨论的程度。

and discuss it in earnest.

这顿饭，

This meal

是这个时代的一个缩影。

is a microcosm of this era.

节目的最后，

At the end of the show,

我想和你聊聊，

I want to talk with you about

这件事背后，那个更根本的问题。

the more fundamental question behind this.

人类历史上，

In human history,

每一种足够强大的新技术——

every sufficiently powerful new technology —

从火药，到核能——

from gunpowder to nuclear energy —

都会面对同一个追问：

has faced the same question:

它的方向盘，应该握在谁的手里？

in whose hands should its steering wheel be held?

是握在最懂它、

In the hands of those who understand it best,

也最有动力推动它的创造者手里？

and are most driven to push it forward?

那样也许最高效，

That might be most efficient,

但是会不会，缺少制约？

but would it lack restraint?

还是握在更多人、

Or in the hands of more people,

更多国家的手里？

more countries?

那样也许更安全、更公平，

That might be safer and fairer,

但是会不会，太慢、太乱？

but would it be too slow, too chaotic?

AI，

AI

正是我们这一代人，

is exactly the new 'fire'

要面对的那个新的"火"。

our generation must face.

这一次，

This time,

我们能不能，

can we

比过去做得更好一点？

do a little better than before?

所以最后，

So finally,

我把两个问题，留给你。

I'll leave you with two questions.

第一，你信任谁，来为 AI 定规矩？

First, whom do you trust to set the rules for AI?

是科技公司，是政府，

Tech companies, governments,

还是某种全新的、大家一起参与的方式？

or some entirely new way where everyone takes part?

第二，回到你自己。

Second, back to yourself.

在这个 AI 飞速变化的时代，

In this era of AI changing so fast,

你希望牢牢握在自己手里的，

what is it that you hope

又是什么？

to hold firmly in your own hands?

这两个问题，没有标准答案。

These two questions have no standard answer.

但是它们，值得你慢慢地想。

But they are worth thinking about slowly.