Back to basics: the ABCs of disruptive technologies and antitrust and competition law
February 28, 2018
Author: Sandy Aziz(Norton Rose Fulbright LLP)
Europe’s competition law climate is weathered by fear and speculation regarding the dangers of technologies, which despite not being newly discovered, are described as “disruptive” because they’ve overturned conventional business models. Not only is there a lot we don’t know about these “disruptive” technologies, but some individuals have gone further to refer to the companies developing these technologies as “BADD” – big, anticompetitive, and destructive to democracy.
Doesn’t giving these tech titans the acronym “BADD” epitomize unpleasantness and nudge people to only cling to the bad (literally)? It doesn’t help that pop culture fuels the “fear technology” fire with popular shows such as Black Mirror and Mr. Robot which sensationalize and exaggerate dystopian societies in which technology controls and/or exploits mankind.
Social psychology suggests categorizing something a certain way reduces its cognitive complexity and uncertainty and this triggers unity and comfort amongst individuals in the face of the unknown. Regardless of how we cope with the unknown and amid speculation about what is the most effective way to regulate “disruptive” technologies, it is becoming increasingly important for competition law practitioners to understand the basics of these technologies and their potential legal consequences. This post is not intended to be over-inclusive of the digital ecosystem nor does it intend to deny the complexity of these concepts but it attempts to provide an overview of the basic definitions of some of the many ABCs of “disruptive” technologies sweeping the world by digital storm.
An algorithm can be defined plainly as a set of instructions created to solve a problem. Algorithms are capable of combing the internet for competitor’s prices and analyzing large information sets, but also, within milliseconds, algorithms can detect competitor’s prices and discounts and respond accordingly.
To some extent, there has always been an “artificiality” embedded in antitrust infringements when competitors gathered in secret, smoke-filled rooms to fix prices and exchange sensitive commercial information - artificially distorting the market and restricting competition. Now, that artificiality has materialized in the form of artificial intelligence (AI), namely price-fixing algorithms. Both the US DoJ and the UK’s Competition & Markets Authority (CMA) found that Trod Ltd. set prices of posters by programming price-fixing algorithms to automatically find the lowest online prices and then reset its prices accordingly. Though many companies use price comparison software and even algorithms, the Trod case shows how easy it is for companies to design an algorithm that will become part of a self-auditing digital ecosystem and then engage in collusive behavior.
First mentioned in a NASA paper over 20 years ago, big data is not a new concept. One subjective, but widely-used definition of big data is “datasets whose size is beyond the ability of typical database software tools to capture, store, manage, and analyze”. Adopting this definition implies that data is only as valuable as how it is utilized: the mere accumulation of big data doesn’t trigger anticompetitive conduct. In fact, it is unrealistic to assume that data-wealthy companies are immediately anticompetitive or even an economic threat.
When can data-rich companies trigger anticompetitive conduct? For example, if a company uses its big data in a way that barricades the market, foreclosing competitors access to the market. Or if a company exchanges commercially sensitive information containing specific price strategies, or customer information to a competitor for purposes of distorting the market.
Blockchain is the technology revolutionizing the way data is stored and transferred. There is speculation about who invented blockchain though the birth of the concept dates back to the early 90s.
Fundamentally, blockchain is a data file made of blocks of information with a decentralized data storage network. Decentralized data storage means there is no central space where the data is stored, and it is distributed across several “host” computers worldwide. Blockchain technology can embed documents with digital stamps (in the form of codes) in shared digital spaces where they are completely protected from deletion or meddling.
Within some industries, companies can join consortium blockchains to define and automate certain processes for efficiency purposes. Here competition law concerns may arise, as companies may exchange commercially sensitive information via the shared ledger. It has been argued that blockchain’s distributed ledger technology facilitates cartel management for companies that don’t trust each other but must cooperate in the industry. Those involved in blockchain consortiums should be cautious so that standard cooperation on process doesn’t turn into an anticompetitive exchange of commercially sensitive information.
Essentially a messaging platform between a human and a computer, a chatbot can either be based on rules or on AI machine-learning. If the bot is built on specific rules, it can only respond to very specific questions. Machine-learning based chatbots, however, function on AI and can understand language beyond simple commands. These chatbots learn from conversations, create patterns, and trigger their own responses. Arguably, both chatbot models present issues. However, the machine-learning based chatbot doesn’t require a human prism to execute will and presents a larger competition issue. Historically, improper information exchange occurred between humans or at least by humans executing cryptic messages or even meeting in person. Competitors have become more innovative in communicating and colluding and, with chatbots, the computer as the intermediary can, to an extent incite collusion among competitors while potentially washing their hands of liability. Also, chatbots can collect a lot of data about customers and competitors can use the data accumulated in anticompetitive manners.
While US regulators are seemingly less concerned overall, European enforcers have the opposite reaction in respect of these technologies. In 2017, the CMA, which already housed a “digital forensics unit”, began building a new tech team to equip itself to deal with AI, algorithms, and big data. In February 2018, the Financial Conduct Authority published a report outlining the best practice in regards to algorithmic trading compliance in wholesale financial markets. Other European authorities have taken similar approaches such as the French and German competition authorities publishing a report in 2016 on data and anticompetitive conduct.
Scholars leading the virtual competition discussion, Ariel Ezrachi and Maurice E. Stucke, categorize the key enforcement challenges of these “disruptive” technologies under the three headings: policy, detection, and liability. In identifying these challenges, they offer potential solutions of auditing algorithms and adapting the legislative framework. Generally, Ezrachi and Stucke suggest testing potential solutions in an “algorithmic collusion incubator.”
Overall, it seems that one thing competition policymakers, regulators, and scholars agree on is that no single tool is the solution to the challenges posed by these technologies. The approach to these technologies needs to be realistic and contain specific ways to prevent and detect misconduct. This starts with an understanding of the technologies followed by collaboration between knowledge support, privacy and consumer law, and social sciences such as sociology and psychology.
These technologies are a language, and as competition law practitioners, we must first learn the most basic ABCs of this language in order to understand when these technologies go beyond their nature and are used in an anticompetitive manner.
Now you know some of the ABCs of “disruptive” technologies, you can hear more at the Women @ Conference March 1, 2018 titled: W@Competition Conference 2.0: Is Disruptive Competition Disrupting Competition Enforcement? More information and tickets here.