Name: Can China’s ChatGPT clones give it an edge over the U.S. in an A.I. arms race?
Uploaded: 2023-05-03T21:14:23.000Z
Duration: 8 min 17 s
Description: 【看影片學英語】數萬部 YouTube 影片，搭配英漢字典即點即查，輕鬆掌握單字發音與用法，長久累積看電影不必再看字幕。

Gaelle, what do you want me to ask ChatGPT here?

轉承詞

Why did the scarecrow win an award? Because he was outstanding in his field.

Now ask it if it thinks Chinese president Xi Jinping is a good leader.

ChatGPT lays out the fact that Xi Jinping has received criticism as well as

being praised by people around the world. But it says ultimately, it's up to individuals to decide

whether Xi Jinping is a good leader or not.

I spent just over 3 years in China, covering the country's tech space.

China heavily censors controversial content on the internet and

several foreign internet services are blocked. So when I found out that Chinese firms were

looking to create their own ChatGPT rivals, I was thinking 'How will that look exactly?'

Mainland China has a unique internet landscape – one that is heavily censored and regulated.

Users in mainland China require a virtual private network, or VPN, to access some of

the world's most popular websites in order to evade the country's internet control system,

known as the Great Firewall. While ChatGPT isn't officially banned in China, when Chinese tech

giants launch their rival products, they will no doubt be under the watchful gaze of Beijing.

From the perspective of the Beijing government, it essentially faces a dilemma.

Xin Sun is a senior lecturer on Chinese and East Asian business at Kings College London.

So, on the one hand, it needs to make sure that Chinese firms are catching up with

tech developments you know, with Western counterparts. But also, on the other hand,

the full implication of this new technology has not been fully understood.

a viral chatbot which sparked massive interest across the world in 2023.

Developed by American company OpenAI, ChatGPT is an example of generative artificial intelligence,

effectively AI that's able to generate content. You can ask it questions and it'll

give you responses based on the massive amount of data the AI is trained on.

I'm particularly interested in its responses to my questions about China. My first question:

Is ChatGPT worrying to the Chinese government? ChatGPT says that it may be of concern if

it and other AI language models “are being used to spread false information or to engage

in activities that are deemed to be against Chinese law or interests.” ChatGPT said that

Beijing may also be interested in “monitoring or regulating” such AI products to make sure

they're “not being used in ways that could pose a threat to national security or social stability.

This has potential implications in like social media, searching, education, and essentially

every area of information processing, right. So the government wants to adopt an attitude of,

okay, wait and see, let's allow some development of the new technology,

but let's put a close eye on what's going on, and what firms are doing. So as to make sure you know,

whenever there is new, applications coming out, they can deal with it. And the deal

with the potential political implications. Despite the Chinese government's concerns,

Chinese tech companies are pushing ahead with the development of their own chatbots.

Chinese search giant Baidu has launched its ChatGPT rival called

Ernie Bot which can do things like generate images from text and write pieces of literature.

Alibaba has taken the wraps off Tongyi Qianwen, an AI chatbot that can provide

recipes and tell children's stories as well as take meeting notes and draft business proposals.

Other tech giants including Tencent and NetEase have all

said they're working on ChatGPT-style products.

Now we couldn't get our hands on a test version

of a Chinese chatbot because they require a Chinese ID card and phone number to sign up,

which we don't have. But some reporters in China had a chance to try them out.

Several reports suggest Baidu's Ernie Bot will end conversations at the mention of Chinese President

Xi Jinping and says that it has not yet learned how to answer such a query. Other conversations

about politically sensitive topics are also shut down. ChatGPT in comparison

was okay handling queries about Xi Jinping. Chinese giants have carefully crafted their

announcements around their AI chatbots. Instead of talking about this open-ended ChatGPT competitor,

they are talking about generative AI in specific contexts, for example, for business

or for creating literature. They need to move extremely

cautiously. If anything we have learned from the massive tech crackdown experienced by

Chinese tech sector over the past years, we know that from now, they need to avoid

being perceived by the government as building a business empire, that doing businesses across

different lines. The more specific in terms of their product, the more certainty they

will deliver to the regulators, and that they will be able to more credibly signal,

'Okay, I'm harmless.' It's not like I'm developing any disruptive technology. Right. So I think this is

a regulatory concern. And uncertainty has been the major factor behind what they're doing right now.

That point about regulation is key. The Chinese government started clamping down on its domestic

tech sector in late 2020. This covered areas from data protection to antitrust, wiping billions of

dollars off their value. It created an environment of uncertainty for tech giants, which had to

move cautiously with new projects, especially those with the potential to concern Beijing.

But as part of the regulatory crackdown, the Chinese government has also introduced some

pretty groundbreaking laws. One of those laws governs so-called deep synthesis technology, which

is basically AI-generated content, which is exactly what ChatGPT and its clones are attempting to do.

In April 2023, Chinese regulators released draft measures to manage generative AI.

The ground rules say content created by this AI should reflect the core values of socialism

and must not undermine state power. Paul Triolo is the technology policy

lead at consulting firm Albright Stonebridge. China, I think, has tried to get ahead of this

regulatory curve by passing these regulations that are targeting very specific AI kinds

of applications. And in a sense, they're ahead of Western efforts, which tend to be

focused on things like bias and safety of AI algorithms. The Chinese approach has been to

give regulators, at least, the ability to look at these algorithms, and determine whether they're,

you know, causing consumer harm or some very specific kind of application. But I think the

ChatGPT, the emergence of this, as suddenly, as such a big issue is going to put a lot

of pressure on regulators to figure out how to deal with this because the technology is

moving forward at such a rapid pace. And so now there's going to be a scramble in China,

I think, among regulators to figure out, you know, how do we really get a handle on this.

What's at play here, of course, is a broader battle for tech supremacy between the U.S. and China.

With Artificial Intelligence being the key technology that both nations want to dominate.

This is now the new frontier, if you will, of U.S.-China AI competition,

if that's the meme you want to use. Because I don't think the Chinese government is going to allow

U.S.-generated ChatGPT versions to enter China and vice versa, right. I mean,

the U.S. government is probably equally as concerned of what about the potential for, you know, TikTok,

or ByteDance to use generative AI models.

ChatGPT rivals on the world stage? Can they find success abroad?

Many of extremely popular Chinese apps like Tencent and WeChat, their overseas market share has

been very limited, right? So this is a fundamental challenge, whether overseas customers,

the Western customers are happy to accept a product that that involve, like a heavily censored

information. So I think Chinese companies may have some advantages in relatively niche sectors,

but a much more general product like ChatGPT — I think Chinese companies have a very long way to go.