Reinventing Capitalism in the Age of Big Data Hardcover by Viktor Mayer-Schönberger and Thomas Ramge
Schönberger is a professor at Oxford and Ramge is a technology writer who has by-lines in The Economist. This book is not their first dance with Big Data. Their objective was to unveil its possibilities to those of us not close to Big Data and Artificial Intelligence (AI), also called machine learning and other terms. Their work sparks ideas, even for readers who do not agree with them.
The heart of artificial intelligence is learning by detecting patterns and matching them with other patterns. It’s what Amazon does when it informs you that, “people who bought X also like Y and Z.” Pattern detecting and matching are the heart of the business models of Amazon, Google, Facebook, Netflix, E-Bay, Spotify, and Apple Music, well known for ripping up competitors’ old business models. They are now wedging storefront retailers out of their markets, just like music distributors before them.
AI becomes “eerie” when applied to observable personality characteristics. For example, Alistair Shepherd discovered how to predict the performance of teams by detecting whether observable personality patterns of team members matched or clashed. Prior experience was much less a predictor than whether team personality patterns meshed. On-line romantic match making services also began to have better success rates when they matched based on reported personality traits.
Pattern matching with AI is advancing rapidly. Pioneers are disrupting markets now. What will happen when preference matching “can be done by anybody?” The nature of markets will drastically change, Reinvent Capitalism that’s what. We might not even call it capitalism, because the important factors will cease to be money and capital. Instead, data and what can be done with it will become dominant.
Companies will drastically change. Spotify is an example. It’s semi-structured, organized into typical agile software teams with very little hierarchy. Companies will find economy of scale failing them; top down control failing them. They have to reorganize for rapid lateral communication, cooperation, and action learning.
Amazon doesn’t work that way today. Jeff Bezos controls top-down what has been described as a data-driven tyranny. Employees monitored in detail like robots call themselves “Amabots.” Schönberger and Ramge aver that Amazon cannot survive this form of organization much longer.
Big Data and AI should augment decision makers’ intuition, not totally displace it. People should see what they otherwise would miss, either because they are not looking for it, or because they are overwhelmed by information. To use AI and Big Data wisely, decision makers should understand what it does; therefore what they should do.
Vision of a New System
What would an economy after money look like? Schönberger and Ramge project that future markets will be data rich, filling in much of the information neglected by price-driven market exchanges. Price is only one data point among many. Instead, invest in a company based on its pattern of functions, not whether its stock is “over or under valued.” No more project bids going for the lowest price; instead match contractor profiles with project profiles. Gone will be the gamesmanship of over-promising, low-balling, and reverse auctioning.
Big Data is brewing a perfect storm in banking. The “spread” between interest charged and interest paid has been under pressure for 40 years. Back then banks could pay most operating expense out of the spread. Now they have to tack on service fees and engage in risky investing. Service fees invite disruptors like PayPal and Apple Pay. These services displace the role of a bank as a trusted intermediary validating transactions. Crowdfunding services keep improving, expanding. Some day peer-to-peer lending will crumble a second big function of banks.
More generally, data rich exchanges will displace capital as the key ingredient of the system. Unicorn companies now thriving on Big Data don’t need much capital or huge numbers of employees. Schönberger and Ramge project that data exchange will displace money as the key medium of exchange.
In addition, pattern matching will displace almost any clerical work that involves pattern matching, in whole or in part. Approving insurance claims is already an example. And faster matching with AI will diminish the lead times of many bureaucratic approvals.
Economic disruption will get worse. Its downsides must be addressed. A big one is that Big Data can be used to bias perceptions, and the book was finished before the Facebook-Cambridge Analytica uproar broke. People may be influenced to move along in decentralized fashion toward a common goal. They can be subtly manipulated to hate each other. Or they can simply be scammed. Systemic feedback loops may deceive us that we are free while manipulating us like robots.
Schönberger and Ramge’s overly rosy view of Big Data potential may blind them to how it can be abused to create Big Brothers. User awareness of what data a service is tracking does not preclude this. They underestimate the education of users to beware of scams, much less the robotizing effect of being funneled deeper into existing preferences, while information outside them is withheld.
But Schönberger and Ramge are aware of the dangers of data monopolies. Facebook and Google eat 90% of all on-line advertising because ads are targeted and viewers can click through to place orders. They’ve been described as too necessary to fail. The authors propose requiring dominant AI services to give data to others that can also use it. They toy with taxing robots or taxing data, assets where “the real value lies.” Is that enough? Being market focused, they largely ignore abuse of AI by natural monopolies like government agencies, or by captive services like health care and education.
I did not find what I was hoping for in Reinventing Capitalism. Can Big Data help us balance multiple objectives when environmental issues must take priority? If Big Data can help match people on teams, it can be trained to help match what we can do to the regenerative needs of a local environment – or can it?
Schönberger and Ramge have mixed ideas about the future of employment when Big Data is widely deployed. They see pattern matching already being used to better match people with work to be done. Matching employees with employers may devalue the role of pay in job selection because if people really enjoy what they do and who they work with, salary is a secondary consideration – as long as it is adequate.
However, masses of people excluded from any meaningful work is a strong possibility, so Guaranteed Basic Income is apt to become a necessary public policy. Funding it might require new thinking about a different economic system because with the present system it would eat the majority of today’s government budget.
Schönberger and Ramge end upbeat. The job of AI and Big Data is to let humans be human, not to prod us into mindless consumption or trick us into pseudo rational living. They do provoke thought, so perhaps they can provoke themselves into thinking more broadly and deeply on how we can overcome downsides to get from here to a dimply perceived there. This reviewer certainly empathizes with that challenge.