China’s plan for domination in Artificial Intelligence

Oct 4, 2018
Vote on Hacker News

AI and bias

 

The already pervasive influence of Artificial Intelligence in our lives

The key selling point of this year’s iPhone, Apple’s most expensive device yet, is not it’s battery life or speed, but the AI capabilities that are built into it. 

One feature, revealed in Cupertino at an early September presentation, was the enhanced camera present in this year’s iPhone. Improvements in the images it gathers and collates have come this time, not just from an improved quantity and quality of the lens(es) on the back of the product, but in the Artificial Intelligence capabilities which are built into it. 

This year’s iPhone no longer takes pictures, it collates data from the lens – which the phone uses to create great looking images, using AI to help produce the best result. It’s an important distinction and just one reason why Apple has now included a ‘Bionic Processor’ in the phone – a CPU with Neural Processing capability.

Apple is just one example of how AI has already started to contribute to the many of the tasks we perform each day. AI is the heart of the Google Searches we do, the smarts behind the shows Netflix recommends you watch and the nudge you need to add something that ‘people like you’ bought, to your cart, when you’re shopping on Amazon. 

AI is also at the core of one of the major behind the scenes struggles for influence in the world. And make no mistake, China is winning the fight. The commercial opportunity is huge. (McKinsey estimate it could generate $13 trillion per year by 2030.) The strategic and military applications are likely to be equally substantial. It’s the modern day space race. 

It’s worth thinking – if, as seems likely, China wins the war over ownership of the most used AI. The impact may be far more reaching than who makes the most money.

China is already leading in the war over AI

This week China announced what appeared to be a softening of its previously extremely hard line on global domination in the area of Artificial Intelligence (AI). China is out investing, our strategizing and out researching America when it comes to Artificial Intelligence. They have a broad, well considered, holistic approach to the topic, which combines 3 areas.

  • AI Strategy: China has a single, clearly stated, National vision around owning the AI market. They have aligned resources from University Educational programs in AI to allowing Alibaba, their biggest eCommerce concern, to run an entire city for them. 
  • New economic model: China ensures that the messy world of capitalism does not undermine the national interest. To this end, Government representatives sit on the boards of China’s major technology conglomerates, to ensure strategy in the interests of the company doesn’t waver too far from the interests of the country.
  • Research and Development: Of course, Alibaba’s AI services are all cloud-based. Alibaba will spend slightly over $7bn USD in Research and Development projects, in the year ahead, making them third in the world behind Alphabet and Amazon but ahead of Ford, IBM and Facebook. They are only a part of the country’s $400+ bn spent since the turn of the century in R&D as a whole. The US still spends more on R&D but China is growing their investment dollars faster and will soon overtake.
  • More data: China simply has more data than anyone else – and data is the most important part of any AI solution. The country contains 1.3bn people and a government which uses advanced surveillance techniques to follow its people. Nearly 25% of the world’s population lives in China and they generate a great deal of information about their behavior and movements.

China also has a strategic advantage in terms of the number of qualified workers to contribute to AI projects. Some estimates suggest that there are only 10,000 suitably trained individuals in the field of Artificial Intelligence, on the globe at this time.

In contrast, the major competitor in the AI saga, America, has little cohesive strategy and appears to be leaving it to the market to make America best in the world at AI. Given their experience with Google, Facebook, Amazon, Microsoft and the rest, who have achieved their own brand of domination, when left to their own devices, that may not seem like such a ridiculous stance to the Trump Administration.

What is ‘Bias’ built into Artificial Intelligence?

There are a number of ramifications likely to follow China’s successful acquisition of the number 1 spot in ownership of Artificial Intelligence. There is, however, one more harder to see, potentially negative outcome which people in the Western World might experience; Bias in favor of a Chinese perspective. 

Kay Firth-Butterfield and the World Economic Forum believes such bias, baked into AI algorithms, either intentionally, or more likely tacitly, could be a real issue, unless we choose to address it head-on.

 

 

In theory, and on the face of it, computers can’t be biased in the solutions they offer. They have no emotional components, no subconscious fears or favors to project on to the data they analyze 

When AI systems output results which contain the prejudice of their designers, that’s Bias in AI.

The problem comes when humans are involved in the design of the algorithms which interpret those data, imposing their own views on it. When AI systems output results which contain the prejudice of their designers, that’s Bias in AI.

In part, at least, the problem is that those designing the software in The West, at Google and Facebook and the other technology behemoths, are typically young white men – at least for now. 

The result of this single gender, single ethnicity, has been some ‘obvious problems’ with some Artificial Intelligence solutions which have been brought to the fore, so far. 

One good example is Google’s AI software which ‘recognized’ African Americans as ‘gorillas.’ 

In simple terms, if your intent, as a programmer, is derived from different beliefs and mental models, to mine, you’ll likely design a different AI engine and get a different AI answer than if I did the same thing. 

Take a simple, possible, example. Imagine you’re designing an AI system to mark an essay submitted by schoolchildren for their High School finals. The title of the paper is ‘How my family contributed to my success.’ There are two AI systems doing the marking. One is Chinese and comes from a cultural heritage which values community and family bonds, one is from Google which favors the rights of the individual and an entrepreneurial approach. 

It’s easy to see the potential for very different content (in the form of essay answers) to be ‘rewarded’ with a higher mark. 

Now, apply that reward thinking and the effect it would have on student behavior, to the entire educational process. For the whole time, they’re at school, the ‘right’ result from their AI marker is to provide solutions consistent with different values. It’s more than possible that kids trained by an AI system, developed by a different cultural framework in China, would behave differently from one designed in the US.

Summing up Chinese bias in AI

Given what are reasonable concerns about bias, built into AI, in existing AI systems, there may be cause for concern as China progress down the path to owning the field.

 In the event that China won the AI war, the effects on our own culture and norms could be more dramatic than we think. The bias could be ‘against’ Western values and impose aspects of Chinese culture on our behavior.

Is such a thing so unlikely? Consider Hollywood, California. It is, unquestionably, the world center for the manufacture of movies and has, since the invention of ‘talkies’ produced and disseminated hundreds of films per year. When asked what you consider heroic, you might answer John McClane, the lone wolf who beat the bad guys at Nakatomi Plaza in ‘Die Hard’. 

Equivalently, you could say it was Ariel from The Little Mermaid. With either answer, the US produced the concepts to which you refer. It could be the sort of influence China stands to gain by winning the fight of who owns the best AI solutions.

 

 

Ralf is an I.T expert and a technology blogger, he writes about mobile phones and the latest technology news.  He currently works at Whatphone.com.au as a content manager and his writings can be seen on various technology blogs. He also loves taking pictures when free. You can follow him on Twitter @IamRalf12.