How do you throw the book at an algorithm? | John Naughton

Policing the internet in the face of big data is proving to be an almost insurmountable problem

W hen, in the mid-1990s, the world wide web transformed the internet from a geek playground into a global marketplace, I once had an image of seeing two elderly gentlemen dancing delightedly in that part of heaven reserved for political philosophers. Their names: Adam Smith and Friedrich Hayek.

Why were they celebrating? Because they saw in the internet a technology that would validate their most treasured beliefs. Smith saw vigorous competition as the benevolent “invisible hand” that ensured individuals’ efforts to pursue their own interests could benefit society more than if they were actually trying to achieve that end. Hayek foresaw the potential of the internet to turn almost any set of transactions into a marketplace as a way of corroborating his belief that price signals communicated via open markets were the optimum way for individuals to co-ordinate their activities.

In the 1990s, those rosy views of the online world made sense. The technology provided a greenfield site for free-market economics. Barriers to entry were low. Information inequalities between buyers and sellers were being reduced by the rise of search engines and price-comparison sites: no more “markets for lemons”. Competition was fierce because a better price was always just a click away. Online markets were becoming, in the jargon of the day, “frictionless”. Apart from Microsoft (which wasn’t in the e-commerce business), there were no industrial giants to exercise monopoly power. And online companies knew relatively little about their customers.

Spool forward two decades and the only thing that hasn’t changed is the evangelical rhetoric of the tech industry. The online economy has been utterly transformed. It’s dominated by huge companies that vacuum up the digital footprints of all their customers and feed them into algorithms that determine prices, respond instantly to competitors’ prices and decide what should be offered to each customer. But the rhetoric of perfect competition, sovereign consumer, free markets and the dangers of government regulation remains locked in the era of Hayek, if not of Smith.

Enter Ariel Ezrachi and Maurice Stucke, two distinguished scholars specialising in competition law, who decided to ask if the online emperor has any clothes. Is the veneer of competitiveness provided by the vigour and diversity of our online marketplaces just an illusion? Their book – Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy – provides some sobering answers.

Traditional competition law is about firms and their activities. The great insight underpinning Ezrachi’s and Stucke’s book is that, in a digital world, competition law will be mostly about algorithms and big data because these are the forces that now determine what happens in online marketplaces. The book focuses on three particular areas in which anticompetitive and manipulative behaviour is possible and, in some cases, already evident. They are: algorithmically enabled collusion (essentially computerised price fixing); behavioural discrimination (the process by which different customers get different offers depending on their data trails); and what the authors describe as “the dynamic interplay among frenemies” (ie the new ways in which firms can be both collaborators and competitors).

In each of these areas, Ezrachi and Stucke dig deep into the ways in which algorithmic and big-data analytics combine to produce behaviours and outcomes that are – or could be – troubling for society. They then go on to discuss the extent to which existing competition law and legal precedents may – or may not – be able to address abuses. In the analogue world, for example, the law can deal with tacit collusion on price fixing because corporate executives are the agents who do it and it is possible often to prove intent. But Ezrachi and Stucke come up with plausible scenarios in which the pricing algorithms of rival firms may produce outcomes that are indistinguishable from tacit collusion, yet difficult to prosecute. Who do you sue when the “culprit” is a machine-learning algorithm? And how would you prove intent when an algorithm produces outcomes that its programmers could not have predicted?

Even in the analogue world, competition law was ferociously complex and riddled with ideology: witness the way market dominance morphed, under the influence of the Chicago Law School, from a social evil into proof of industrial excellence. In the networked world, legal regulation will be orders of magnitude more complicated. In fact, one of the effects of having spent a week with this book is a troubling uncertainty. Ezrachi and Stucke have made a convincing case for the need to rethink competition law to cope with algorithmic capitalism’s potential for malfeasance. The bigger question, though, is whether, given the mind-bending complexity of the technology, we are capable of mastering it.

Contributor

John Naughton

The GuardianTramp

Related Content

Article image
The tech giants, the US and the Chinese spy chips that never were… or were they? | John Naughton
A sensational Bloomberg story about a major hardware hack was swiftly denied. But the journalists aren’t backing down

John Naughton

13, Oct, 2018 @4:00 PM

Article image
How Amazon puts misinformation on your reading list | John Naughton
Algorithms routinely come up with ‘recommendations’ for anti-vax ‘bestsellers’ or juices that can cure cancer

John Naughton

08, Aug, 2020 @4:00 PM

Article image
Let's not forget, Bill Gates hasn't always been the good guy…
A quarter of a century after the launch of Windows 95, it’s striking to remember the philanthropist we know today as a mogul hellbent on world domination

John Naughton

29, Aug, 2020 @3:00 PM

Article image
Machine learning: why we mustn’t be slaves to the algorithm
There’s a huge buzz around the marvels of machine learning, but computers are only as good as the biased data we feed them

John Naughton

16, Oct, 2016 @6:00 AM

Even Google won't be around for ever, let alone Facebook

In the world of internet technology a company can go from zero to hero in a very short time, writes John Naughton

John Naughton

03, Mar, 2013 @12:04 AM

Article image
All I want for 2021 is to see Mark Zuckerberg up in court | John Naughton
The tech giants’ law-free bonanza is coming to an end on both sides of the Atlantic, but let’s speed up the process

John Naughton

02, Jan, 2021 @4:00 PM

Article image
Wanted in the digital monopoly age – powers to curb the hold of online giants | John Naughton
Analysis by Lina Khan reveals how Amazon exploits outdated thinking about competition

John Naughton

16, Sep, 2018 @6:00 AM

Article image
Will blockchain fulfil its democratic promise or will it become a tool of big tech? | John Naughton
Engineers are focused on reducing its carbon footprint, ignoring the governance issues raised by the technology

John Naughton

15, Jan, 2022 @4:00 PM

Article image
When it comes to Amazon, breaking up is hard to do | John Naughton
Given the problems involved in regulating the tech giants, the EU commission’s targeted investigation seems the smartest way to achieve results

John Naughton

14, Nov, 2020 @4:00 PM

Article image
Been duped on Facebook and Amazon’s platforms? Well, you’re not alone | John Naughton
Allowing third parties access to the online giants is as risky as it is profitable

John Naughton

31, Aug, 2019 @3:00 PM