A Blog by Jonathan Low

 

Feb 2, 2022

Why China's Data-Centric AI Strategy May Be Suboptimal

China's primary advantage in AI, aside from the internal talent it is able to generate, is its ability to command data from its @1.45 billion people. 

This gives it the ability to test and generate new models with virtually no restriction. The problem it has, though, is that few, if any, non-Chinese experts are welcome even if they were interested in working there - and that politically-driven oversight decisions may stifle innovation in ways that are not problematic in the west. This, despite its investment in AI could well limit its ability to compete in this field for the future. JL  

Tristan Greene reports in The Next Web:

China‘s main advantage, when it comes to AI, is socialized data, (which) gives the Chinese government carte blanche to develop bespoke systems for its citizens. As with any walled-garden, it’s easy to share the benefits contained within. But the PRC is non-competitive when it comes to luring outside talent. (And) in China, government restrictions prevent its top AI firms from prospering on the global stage in the same ways as Meta, Google, Microsoft, or Apple. The combined resources, scientific contributions, and technological superiority shared by US academic and corporate institutions in AI is enough to overcome the advantages given China by its data.

There are few more compelling story lines at the intersection of Wall Street and Fear Street than China‘s rise to global prominence in the field of artificial intelligence.

You don’t have to look very far to find a military or financial expert who believes China‘s AI program will some day surpass the capabilities of its democratic counterparts in Silicon Valley.

But, as we’ve written before, the idea that China is in second place behind the US is a bit misleading. Sure, it’s technically correct, but we’re talking about a very distant second place. Currently, it would be a huge stretch to call it a race.

 

China‘s main advantage, when it comes to AI, is socialized data. But data can only take a nation so far.

If you’re trying to optimize the national healthcare program, for example, having unfettered access to the private data of billions of citizens is invaluable. However, personal information isn’t exactly useful if you’re training drone swarms to identify and assault soft targets.

Communism

The means of production, in 1949, when the People’s Republic of China was formed, was agriculture and industry. Today, it’s technology.

The advent and implementation of deep learning has lead to a revolution in China‘s healthcare and infrastructure. Since 2018 the PRC’s expanded it’s national insurance to cover nearly every citizen in China.

Doctor and author Eric Topol described the impact AI’s had on that growth in a recent article by Eleanor Olcott for the Financial Times:

China has shifted faster than the US in medical AI from research to implementation, driven in part by the availability of high-quality data, says Topol. “China has a massive data advantage when it comes to medical AI research,” he says, explaining that Chinese researchers can train AI models on data sets covering entire provinces.

In contrast, their US counterparts are restricted to working with information from single hospitals — largely operated by private businesses that keep records on internal servers.

This paints the picture of a disjointed US medical system and, in many regards, it’s true. There’s little hope the US can ever hope to accomplish through capitalism what socialized medicine in a communist system can. However, there’s more to prosperity than free healthcare.

Capitalism

Where China‘s able to convert citizen data into internal action policies, capitalist practices have allowed US AI companies to thrive in ways their counterparts in China and the EU aren’t able to.

In China, government restrictions prevent its top AI firms from prospering on the global stage in the same ways as Meta, Google, Microsoft, or Apple. And the presence of regulations limits the EU market in ways Silicon Valley doesn’t. As icing on the cake, most big tech companies in the US pay little or no taxes. 

This leaves US soil fertile for the development of monolithic business empires with a GDP eclipsing many modernized countries. And, with that prosperity, comes a much easier time recruiting the world’s most talented scientists and developers.

Toss in the fact that a significant portion of cutting-edge academic research in the US and around the globe is funded by DARPA, the Pentagon’s think tank, and the communist advantage is minimized beyond the PRC’s own borders.

The long run

Communism and socialized data gives the Chinese government carte blanche to develop bespoke systems for its citizens. As with any walled-garden, it’s easy to share the benefits contained within. But the PRC is non-competitive when it comes to luring outside talent.

And that goes for corporate buyouts and mergers as well. While Tencent and Baidu may have an enormous presence in the global AI market, there’s no reality where DeepMind gets purchased by a Chinese corp over a US one.

It’s beyond the scope of this article to argue the benefits of AI-powered socialized healthcare versus having big tech run wild with user data as it does in the US.

But any conversation around the “AI race” should be prefaced by the knowledge that the combined resources, scientific contributions, and technological superiority shared by US academic and corporate institutions in the field of AI is more than enough to overcome the advantages given China by its policy of socialized data. 

It’s likely that China‘s approach to AI will result in revolutionary new achievements in its healthcare and infrastructure programs. But, arguably, those are areas where Canada, the Netherlands, Japan, and dozens of other countries also outperform the US.

0 comments:

Post a Comment