Theme: Algorithmic Collusion

Artificial intelligence, as a specific form of innovation, creates a conundrum for legal scholarship. The evolution of artificial intelligence has been proven hard to predict. The only thing what is certain is that it will progress. If predictions are hard to make, it is also hard to assess what the risks are related to artificial intelligence. Without knowing the risks, legislators have no objective basis to design a proper legal framework. Due to the uncertainty, the legislator may rely on the precautionary principle and do nothing. When the legislator realizes that he must intervene, he could be confronted with a Collingridge dilemma. Society may have embraced the new technology and not accept regulatory intervention anymore. Early intervention may have the opposite effect and freeze innovation.


Antitrust scholarship is facing the same problem. Artificial intelligence allows firms to analyze massive amounts of information, generally known as big data. This analysis may be used to facilitate price setting. This price setting will not always be beneficial for consumers. Artificial intelligence could result in collusive price setting, even in non-concentrated markets. The ease with which such a collusion can be established makes these scholars argue for a revision of antitrust law and theory. In the end, this form of collusion resembles tacit collusion and that is not punishable under current antitrust law.


The call for legislative intervention has been questioned. Buyers could use technology to undermine the artificial intelligence of seller. This situation is often called the algorithmic consumer. It is further questioned that algorithms will necessarily lead to collusion. Because of the alternative views, there is a call to avoid panic. Without collecting evidence, antitrust policy should not be disrupted.


Which of the suggested approaches should be taken by the enforcement agencies? Direct evidence that tells us which approach to take does not exist. When we have the evidence, intervention may lag behind events as know-how on artificial intelligence may have moved on. In order not to create the pitfalls described above, alternative regulatory models may have to be developed to promote innovation without jeopardizing the position of consumers. Generally formulated, this conference seeks an answer to how to deal with pricing strategies chosen by artificial intelligence from an antitrust law perspective without knowing the exact risks attached to artificial intelligence. Possible topics could be

Are the presumptions under the current debate, like algorithmic homogeneity, appropriate?

What empirical evidence does exist?

Is price discrimination more likely to happen?

Should we re-conceptualize of tacit collusion?

Should we adjust the rule of reason?

Can we advocate for algorithmic transparency?

Can we test algorithms?

Should we rely on market based solutions?

Is national legislation up to its task?