This chapter examines the algorithmic collusion debate. It observes that four models of algorithmic collusion have been developed (‘Messenger, Hub and Spoke, Predictable Agent and Digital Eye’), which are widely adopted within the academic debate without many critical voices. The actual debate centred on the last two scenarios, which assume the algorithm's ability to automatically reach a colluded outcome. The centre of the debate was on the question of whether algorithms can collude autonomously and how competition law can or should respond. The former question tells the technological debate, whereas the latter recounts the legal debate. This chapter elaborates on the details of both debates. The technological debate focuses on the following issues: what types of algorithms can collude under what conditions; can algorithms end in pricing strategies other than algorithmic collusion; can it be possible that algorithms will never collaborate in a real-world scenario; can audit existing algorithms or testing algorithms prevent anti-competitive conduct. The legal discussion has been marked from the start by a significant divide between those who believe competition law is enough for dealing with algorithms and others who believe competition law must be modified. The call for change can be divided between re-interpreting existing concepts and introducing new concepts. A more drastic option that has been suggested is the creation of regulatory instruments outside the traditional competition law framework. Finally, the chapter introduces the book against the backdrop of the debate.
