Artificial intelligence in China lags behind the West – 03/10/2023 – Igor Patrick

Culinary recipes, academic texts that are almost always inaccurate and even resume writing. The reader most connected with the digital world must have seen these and many other ChatGPT applications circulating around the internet in recent months. For many, the first real interaction with an artificial intelligence, the tool developed by OpenAI that has exploded in popularity, capable of incredible results in multiple languages. Except Chinese.

While discoveries about the site’s possible applications were pumping in the West, internet users on the other side of the world were delighted with the bizarre responses of the chat based on neural language when writing in Chinese. The misconceptions ranged from wrong characters that completely changed the meaning of the answer to very distorted information about basic information about China.

There was then no lack of local companies to surf the wave of popularity of the American company, announcing supposedly more efficient Chinese competitors. Search giant equivalent to Google in the domestic scenario, Baidu was the first to take the lead, announcing the launch of Ernie — in theory trained with more than 100 billion parameters in Chinese since 2019 and practically ready to be incorporated into various tools of the company.

In the academic area, Fudan University in Shanghai even launched a test model named MOSS (a reference to the robot from the Chinese blockbuster “Drifting Earth”), although the application was taken offline due to the boom in interest from Internet users. eager to experience something similar to the wonders of ChatGPT.

So much positive news makes one believe that a possible Chinese leadership in the field of artificial intelligence is a matter of time – but it will not be like that.

Government incentives abound. For some years now, Xi Jinping has placed the intelligence sector as essential in the basket of priority technologies in China. In 2022, Beijing even sent the so-called “Position Paper on Strengthening the Ethical Governance of Artificial Intelligence” to the UN, indicating how it intends to deal with the issue in the diplomatic sphere by defending that “to ensure safe, reliable artificial intelligence and controllable, technology will need to be human-centric and programmed to do good.

Still, Chinese Science and Technology Minister Wang Zhigang seems wary of following the West in the endeavour. During an interview given on the sidelines of the Two Sessions taking place in Beijing, Wang urged caution on the part of Chinese companies and said he believed that a tool like ChatGPT is still something very difficult to achieve in China.

It’s not just about technical expertise, it’s also about political sensitivity. In the past, natural language test models that learned from users have had serious problems — in 2016, for example, a Microsoft AI on Twitter became a Nazi in less than 24 hours by learning from the interactions it received.

The alternative, then, involves feeding technology with data carefully vetted by the country’s programmers and politicians, which is far from being a simple task given the billions of possibilities in a fluid conversation between a machine and a human. What if Chinese AI turns anti-Communist? What if he starts criticizing party leaders? The stakes are very high, and the sensitivity of the Chinese leadership even greater.

It is likely that we will then follow two strands of the development of artificial intelligence. While the US leads the field of natural language such as ChatGPT and Bard, a future competitor developed by Google, China is advanced in what it does best: applicability. With government support and an industry that needs to bet on innovation all the time to stay relevant in a highly competitive market, the technology must be applied to more mature sectors – there is already talk of developing Chinese autonomous cars with artificial intelligence, adaptation for banking systems and enabling the digital economy.

Watching the impact of such a crucial sector for human life in the coming years will therefore be more than a patent war: perhaps the way in which artificial intelligence is used will be the dividing line between opinions in Beijing and in the West about how the technology sector must operate. Difficult to predict at the moment, but the competition promises big news with worldwide impact.

PRESENT LINK: Did you like this text? Subscriber can release five free hits of any link per day. Just click the blue F below.

Leave a Comment