Why AI Is Incredibly Smart and Shockingly Stupid | Yejin Choi | TED

But first, let's get philosophical哲学的 by starting with this quote by Voltaire, an 18th century Enlightenment philosopher, who said, "Common sense is not so common."

但在开始之前，让我们先探讨一下哲学问题，从伏尔泰的一句名言说起：这位 18 世纪启蒙运动哲学家曾说，“常识其实并不常见。”

Turns out this quote couldn't be more relevant to artificial intelligence today.

事实证明，这句名言对当今的人工智能来说再贴切不过了。

Despite that, AI is an undeniably不可否认地 powerful tool, beating the world-class "Go" champion, acing college admission tests and even passing the bar exam.

尽管如此，人工智能无疑是一个强大的工具，它击败了世界级的围棋冠军，在大学入学考试中取得高分，甚至通过了律师资格考试。

I’m a computer scientist of 20 years, and I work on artificial intelligence.

我是一名拥有 20 年经验的计算机科学家，专门从事人工智能工作。

I am here to demystify使……不再神秘 AI.

我的目标是揭开人工智能的神秘面纱。

So AI today is like a Goliath.

所以，当今的人工智能就像歌利亚巨人一样庞大。

It is literally very, very large.

它的规模 literally（字面上）非常、非常大。

It is speculated推测 that the recent ones are trained on tens of thousands of GPUs and a trillion words.

据推测，最近的一些模型是在数万个 GPU 和一万万亿个单词的数据上训练出来的。

Such extreme-scale AI models, often referred to as "large language models," appear to demonstrate展示 sparks of AGI, artificial general intelligence.

这种超大规模的人工智能模型，通常被称为“大型语言模型”，似乎展现出了通用人工智能（AGI）的火花。

Except when it makes small, silly mistakes, which it often does.

除非它犯下一些微小而愚蠢的错误，而这正是它经常做的事。

Many believe that whatever mistakes AI makes today can be easily fixed with brute force, bigger scale and more resources.

许多人认为，无论人工智能今天犯什么错误，都可以通过蛮力、更大的规模和更多的资源轻松解决。

What possibly could go wrong?

究竟会有什么不对劲呢？

So there are three immediate challenges we face already at the societal level.

因此，我们在社会层面已经面临三个紧迫的挑战。

First, extreme-scale AI models are so expensive to train, and only a few tech companies can afford to do so.

首先，超大规模人工智能模型的训练成本极高，只有少数科技公司负担得起。

we already see the concentration集中 of power.

我们已经看到了权力的集中化。

But what's worse for AI safety, we are now at the mercy of those few tech companies because researchers in the larger community do not have the means to truly inspect and

但更糟糕的是，就人工智能安全而言，我们现在受制于那几家科技公司，因为社区中的研究人员没有能力真正去检查并

dissect仔细分析 these models.

剖析这些模型。

And let's not forget their massive carbon footprint and the environmental impact.

别忘了它们巨大的碳足迹及其对环境的影响。

And then there are these additional intellectual questions.

此外，还存在这些额外的智力层面的问题。

Can AI, without robust强健的 common sense, be truly safe for humanity?

如果没有健全的常识，人工智能真的能对人类安全吗？

And is brute-force scale really the only way and even the correct way to teach AI?

而且，依靠蛮力扩大规模真的是唯一的方式，甚至是正确的方式来教导人工智能吗？

So I’m often asked these days whether it's even feasible可行的 to do any meaningful research without extreme-scale compute.

所以，如今我经常被问及，在不具备超大规模计算能力的情况下，是否还能进行任何有意义的研究。

And I work at a university and nonprofit research institute, so I cannot afford a massive GPU farm to create enormous language models.

我在大学和非营利研究机构工作，因此无法负担庞大的 GPU 农场来创建巨型语言模型。

Nevertheless, I believe that there's so much we need to do and can do to make AI sustainable and humanistic人本的.

尽管如此，我相信我们有许多需要去做且能够去做的事情，以使人工智能更具可持续性和人文关怀。

We need to make AI smaller, to democratize使大众化 it.

我们需要让人工智能变得更小，以实现其普及化。

And we need to make AI safer by teaching human norms and values.

我们需要通过教授人类的规范和价值观，让人工智能更安全。

Perhaps we can draw an analogy类比 from "David and Goliath," here, Goliath being the extreme-scale language models, and seek inspiration from an old-time classic, "The Art of War," which tells us, in

也许我们可以从“大卫与歌利亚”的故事中汲取类比：将极端规模的语言模型比作歌利亚，并从《孙子兵法》这一古老经典中寻找灵感，因为书中告诉我们……

my interpretation解读, know your enemy, choose your battles, and innovate your weapons.

依我之见，那就是要知己知彼，选择战场，并革新武器。

Let's start with the first, know your enemy, which means we need to evaluate AI with scrutiny详细审查.

让我们先从第一条开始：知己知彼，这意味着我们需要审慎地评估人工智能。