Networking the global middle class is an end in itself
In the superpower race to build AI, data-pool breadth matters
Currently reading the updated version of Prediction Machines (hat tip Carmi Zlotnik at Legendary), which examines “the simple economics of artificial intelligence.”
The trio of authors’ take on AI is to describe it as a “prediction technology” versus an intelligence unto itself (at this point in its development, because true stand-along intelligence, in their judgment, is decades off). The whole point of better predictive power is reduced uncertainty, allowing us all to “exploit the new economic realities predicated on cheap prediction.”
I really like this straightforward approach because it reduces all the hyperbole about AI (and their associated killer robots!) to something more readily understandable. It also better contextualizes the development of AI within globalization — i.e., as another contributing element to increased certainty in a world that is all too often portrayed as being full of “chaos” (complete BS). Additionally, the characterization of “cheap prediction” goes along with globalization’s many historical enablers (cheap transport, cheap communications/networking). To this cluster of “cheap” I would also add “cheap sensoring” with the rise of the Internet of Things (IoT).
In America’s New Map, I summarize these developments thusly:
We spend a lot of time today debating the issue of censorship: whose views are legitimate for public expression and which views, doing far more harm than good, warrant suppression by authorities. This timeless argument expands exponentially with the pervasive sensoring of our lives (embedding of sensors within everyday objects) made possible by the combined rise of the Internet of Things (the connectivity of all those smart devices) and fifth-generation (5G) broadband cellular networks (delivering digital services ten times faster than 4G). Put simply, the more sensors there are, the greater the potential for censoring speech and behavior.
This is why the US government is so concerned about China’s domination of 5G networks being rapidly constructed around the world today, particularly those involving its national flagship companies Huawei, world-leading provider of 5G “backbone” telecommunications infrastructure, and Hikvision, world-leading provider of video surveillance equipment. Eventually, the IoT will develop into the Artificial Intelligence of Things (AIoT) in which cognitive computing mimics the human brain, allowing AI networks to operate with minimal human intervention and develop on their own.
If that strikes you like the jumping-off point for conscious computers to overthrow their human masters per the Terminator movies, then let me just note that Beijing calls its nationwide video surveillance system “Skynet.” It turns out you really can make this stuff up, and China aims to get there first by being the leading worldwide hardware provider for 5G, the IoT, and the AIoT.
So, per Prediction Machines, what China achieves by networking the Global South and that vast portion (by number) of the majority global middle class, is not just direct access to those consumers, along with mostly passive surveillance of their daily lives (Beijing’s paranoid security angle). It’s also about scaling up the AI data source pool in the race to constantly improve predictive capacity.
If you’re a single-party state governing well over one billion souls, you have a deep and abiding demand for certainty, meaning your primary interest in AI is its capacity to make your rule that much more certain and thus less vulnerable to shocks of any sort.
Again from America’s New Map, we get a sense of Beijing’s ambitions:
China peddles that Orwellian vision around the world to frightened governments eager to control their restive middle class and tech-savvy youth. It is also why Beijing vacuums up Big Data on everyone on this planet to predictively identify those who are, or may become, a threat to Communist Party rule—Minority Report on a global scale.
We’ve all heard the variant about China vacuuming up DNA information globally. The logic there is the same: not just seeking advantages-leading-to-domination in the biotech industry, nor just another method for surveilling society, but ultimately a huge source of big data leading to improved AI-driven prediction machines.
Which gets me back to Prediction Machines, where, in the conclusion, the authors examine the likelihood of one superpower being advantaged over the rest on AI development.
They give three reasons why China might well prevail in this race:
China is “spending billions on AI, including big projects, start-ups, and basic research”
[Not an inherently scary notion to me, because I don’t see technology as being hoard-able in our connected world.]
Sheer scale, as “prediction machines need data, and China has more people to provide that data than anywhere else in the world. It has more factories to train robots, more smartphone users to train consumer products, and more patients to train medical applications.”
[True enough, but that just makes me think India’s rise will soon enough balance out that advantage, in effect, advantaging both superpowers relative to the less populated West (EU, US).]
Better and deeper and wider “data access,” because Beijing isn’t constrained, privacy-wise, on what it collects and brings that same attitude and ambition to its overseas operations.
[Europe is far more concerned with privacy, as is the US, and so both are far more likely to demonize and thus disable, through regulations we all see as necessary, their big tech firms with regard to such capabilities and capacity for data collection. See, for example, dozens of state AGs going after Meta (Facebook and InstaGram) over their purposefully addicting techniques regarding pre-teens and teenagers.]
The book’s authors naturally fret about a “race to the bottom as countries compete to relax privacy restrictions to improve their AI position,” but I spot one missing element in their thinking: the rise of a majority global middle class increasingly centered in emerging/rising economic powers (think about a Nigeria, for example). There, we’re typically not talking about the same Western liberal, pluralistic approach on privacy issues, which is why, I imagine, Beijing has to view the Global South in general as such a target-rich environment.
I may well be spinning my wheels on various points here, as I am not an AI expert. But I think my basic logic here holds in the sense that all I’m doing in this analysis is identifying yet another reason why China views the Global South not as the “problem” (the Western view, particularly when one addresses climate change) but as the “prize” targeted by its massive offerings (Belt and Road Initiative, Smart/Safe City surveillance 5G package, etc.).
It’s not just about China maintaining access to resources/energy.
It’s not just about tapping sequentially-unfolding demographic dividends (SE Asia, South Asia, SW Asia, Africa) to access that cheap labor.
It’s not just about gaining access to rising middle classes (and their consumption) that accompany the combination of a successfully cashed-in demographic dividend and deep integration into global value chains.
On top of all those things, it culminates with a data-aggregation grand strategy designed to provide the Chinese government the most powerful prediction machinery known to humanity. That gives them an edge on promulgating new rulesets throughout globalization and, as those rule proliferate, it provides Beijing with some genuine international authority on the evolution of global institutions.
Structure and norms: if you want to re-order the system you must address/transform both, and AI could well be China’s path to that outcome. After all, most of their brand sale around the world is about their relative stability/certainty versus our fundamentally erratic behavior.
You know the old bit about how the best way to predict the future is to invent it?
Well, AI means that logic sort of runs both ways, as in, the best way to invent the future is to predict it with unmatched accuracy.
My book is all about exploring huge tectonic shifts that compel North-South integration across this century. What Prediction Machines gives my thinking is yet another enabler/goal in the form of AI development.
If the future of globalization arrives in the form of a majority global middle class, then accessing and exploiting the predictive data sources/pool that this vast cohort naturally provides is a key aspect of America’s superpower competition with China, India, and others. In other words, enabling the Global South’s rise (and keeping it sound amidst climate change’s impeding decimation of its lands) is not an obligation but an opportunity.
That’s what I call putting another boxcar on the logic train that is America’s New Map.
Thanks again, Carmi!