Much of the current hysteria about the technology industry is due to its highly ambiguous relationship with its users. Driven by the logics of both compassion and indifference, this relationship has always been erratic yet functional. These two clashing rationales, for example, allowed technology companies, frequently painted as Dr Evil, to claim the mantle of Mother Theresa. However, as the unresolved contradictions of these logics pile up, we can’t fail to notice the incoherence of the industry’s overall social vision.
The compassion story has some truth to it. Tech giants have pegged their business models on our ability to consume. Thus, their interests are somewhat aligned with ours: we need a paycheque to buy what’s being advertised. A charitable comparison might be to Henry Ford paying his workers enough to buy his cars; a less charitable might be to slave owners keeping slaves fed not to lose them to exhaustion. However, unlike Ford or slave owners, our tech moguls want someone else to fund their preferred solutions (eg the universal basic income).
Now, the second logic driving these firms – that of indifference to their users – stems from dynamics of competition in this sector. While big tech firms – widely denounced as monopolists – dominate particular niches from search to online shopping, they increasingly compete at the higher level of data services. Thus, they follow each other into new territory, from cloud computing to self-driving cars. For many – consider Amazon – such information services already deliver higher profit margins than traditional activities.
The secret sauce here is artificial intelligence (AI) and AI’s own secret sauce is the data harvested from users during earlier, consumer-oriented stages of development. Once built, this AI capacity can be lucratively rented out to governments and companies. While boosting profits, this shift to services would also leave individual users, the former darlings of the tech industry, without anyone to subsidise their incurable cat fetish. Worse, there’d be no one in sight to pick up that cheque.
Today’s digital economy is not what it seems – and tech giants know this. As Andrew Ng, the founder of Google Brain project and then the head of AI at Baidu, the Chinese search giant, said in a January 2017 talk at Stanford, “at large [tech] companies, we often launch products not for the revenue but for the data … and we monetise the data through a different product.”
There is hardly a better illustration of what I call “data extractivism”. To see its logic in action, look no further than Google. Recently, it has launched Cloud AutoML, a service that helps businesses tap into its machine learning infrastructure (built with the data its user generated) to train and build their own AI models – much of it for a fee, of course.
On the other hand, being nice to users – giving them free funky features to find art that matches their faces, as Google has recently done with its Arts & Culture app – still pays off, as it helps to fine tune existing AI. But for how long will tech firms need us to train them?
The economics of data extractivism suggests that it won’t continue forever – it will stop once AI, trained with all that extracted data, works well enough. The future, in other words, belongs to the stingy, fee-charging Google of Cloud AutoML, not the generous, fee-waving Google of Arts & Culture.
Our defining analogy of the digital economy, thus, might be wrong: data is not the new oil – its importance might be short-lived – but AI probably is. And, with its rise, the tech sector becomes the truly indispensable, too-big-to-fail industry. It’s one thing for tech firms to worry about us affording a pair of trainers. It’s another thing for them to hold monopolistic access to precious AI-based services needed by, well, almost everyone.
Consider fights against fake news, cyber attacks, cancer: AI is deployed in all these efforts. The world could easily survive the demise of providers of digital advertising but it won’t dispense – not today – with AI solutions to its crises. Just ask Emmanuel Macron or Justin Trudeau, two ambitious leaders who are competing to lure Google or Facebook to expand AI operations in their countries.
What, then, can we make of Mark Zuckerberg’s latest pledge to fix his platform, ensuring “time spent on Facebook is time well spent”? Since it came after several Silicon Valley engineers confessed to helping fuel user-addiction, one can infer the general direction of this new and revamped Facebook 2.0. Tapping into the rhetoric of compassion, it will promise to eliminate junk content, deploying AI to find life-enriching posts. It will also say that the more it knows about us, the better its recommendations would be.
But pause for a second. How did we end up in a world where tech firms build addictive services to grab our data – only to pitch AI solutions for the very addiction problem they created? And what happens once, under the pressure of competition and considering the looming end of data extractivism, the logic of indifference overtakes the logic of compassion? And why would Facebook, once it no longer needs our data, offer its “mental yoga-as-a-service” offering for free?
The likely coda of the digital economy is not pretty: addicted to junk content and losing our way in the infinite memes of dubious provenance, we, the online surplus population, will be eventually asked to fend for ourselves. The tech firms will surely have superb AI protection to sell us. The cognitive elites will prosper, fasting on the digital equivalent of kale and quinoa and browsing the artisanal, handcrafted content hidden from the uninitiated.
The rest will be gorging on cheap, trivial, AI-generated memes – until, at least, we buy the premium package of our favourite platform and regain some sanity. Money spent on Facebook will be money well spent.
via Artificial intelligence (AI) | The Guardian http://bit.ly/2iFrAme
January 27, 2018 at 10:09PM