Skip to Main Content

Artificial Intelligence and Antitrust Law

June 6, 2017

Computer Screen with Code

Artificial intelligence (AI) goes beyond previously existing pricing algorithms that relied on static, preset formulas based on fixed inputs. Now, AI has the ability to process data and communicate with other computers well beyond the capability of human actors. These new technologies present unique challenges as they intersect with antitrust regulation and merger law. Questions now arise as to whether current antitrust regulatory schemes will be sufficient to meet these new needs and concerns.

New Antitrust Challenges Posed by Artificial Intelligence 

Traditional antitrust law is based largely on regulatory schemes that applied to older industries like oil and steel. Regulations tended to focus on broad factors like the size of companies; however, in the area of big data, concerns are now focused on the capabilities and potential of specific tools like AI technology. The nature of these programs and algorithms is producing new challenges. For instance, AI-based pricing tools are now able to:

  • Comb the internet for competitor prices
  • Search databases for relevant historical data
  • Analyze digitized information sets
  • Arrive at pricing determinations and solutions

All of this can occur in milliseconds, which means that data is now being generated at unprecedented rates. Theoretically, this should result in wider consumer choice and more competition. However, conflicts can arise when artificial intelligence is explicitly designed to facilitate collusion with other AI or computer programs, or when it is designed to parallel pricing changes by competitors. 

What sets AI technology apart is its ability to communicate independently of humans and interact with other AI programs. This can become problematic, as existing antitrust laws often take into account human intent and action, and even emotional fears. In contrast, automated pricing bots will not respond to various incentives or deterring factors. Also, robots do not leave behind trails of email communications or voicemails that can be used as evidence in court. These elements will make it more difficult to detect competition issues over time.

Concerns Regarding Mergers and Collusion

In particular, AI technology has created anti-competition discussion regarding two main aspects: mergers and collusion.

Merger Concerns

The sheer amount of data involved in merger and acquisition situations could give companies unfair advantages against competitors. Some are concerned that the huge data sets owned by tech giants might act as barriers to entry for smaller startups, thus limiting competition in some fields. An example of this is Facebook’s acquisition of WhatsApp, which raised concerns that the social network may increase its market power through additional data. Facebook has also been fined $122 million due to a misleading information issue in the said acquisition. 

Others scholars argue that possession of large amounts of data does not necessarily constitute an unfair competitive advantage, given the vibrant market for the collection and sale of data. In addition, new technologies are making it easier for smaller market entrants to obtain, store, and analyze the data they need to be competitive. Also, as many of these data companies offer free services, authorities may have difficulty evaluating other issues such as privacy and consumer protection. 


One of the biggest concerns with artificial intelligence is the prospect of the technology being used to facilitate collusion and parallel pricing. As pricing systems become increasingly automated, sophisticated algorithms may render face-to-face collusion between people obsolete. Instead, collusion may occur as a matter of course, as AI systems communicate and coordinate with one another. 

A Harvard Business Review article outlines four distinct ways in which sophisticated computer algorithms can be used to achieve collusion, possibly leading to longer-lasting pricing cartels. These have varying levels of human involvement. As human control over programs becomes less and less, it can become more difficult to prove intent to coordinate prices. Collusion situations might look like the following: 

  • Humans collude in person and use the programs and technology as a tool to achieve pricing aims
  • One algorithm emerges as an industry influencer, acting as a “hub” around which industry-wide collusion revolves
  • There is no longer any formal agreement between competitors. Instead, algorithms exist in parallel where they continually adjust against each other’s data and market prices (tacit collusion)
  • “Digital eye:” technology advances so far in processing high volumes of data that a bird’s eye view of the market is achieved; also, the autonomous decision-making and learning of AI may create collusion that is drastically more difficult to detect

Thus, regulation becomes increasingly complex and difficult as the human actors withdraw more and more into the background.

Another issue connected with AI and price collusion algorithms is that it may be difficult to assess the effects on shareholders, who may be in the dark when it comes to the existence of such algorithms. As long as transparency is an issue, disclosure of potential risks to shareholders will also be difficult, especially given the complexity of such artificial intelligence programs. Algorithms are able to share information so quickly and make such immediate adjustments and decisions that shareholders may not be aware of changes. 

Moving Forward 

Antitrust authorities have traditionally relied on indicators such as size to determine when they should intervene. The nature of AI technology will necessitate a radical rethinking of antitrust regulation approaches. Now, regulators will have to take into account other factors when analyzing the impact of merger deals, like the extent and reach of a firm’s data assets, and the nature of the data usage (not just the quantity of data involved).

Another suggested approach is to increase the transparency of certain kinds of data (with, of course, users’ consent), and in some cases, to mandate the sharing of various types of data.  Other approaches include the creation of a shared set of AI principles for companies to follow, and even the creation of a neutral, non-governmental organization to oversee activity and investigate conflicts. 

One thing is certain — adjustments need to be made quickly, as issues regarding data and AI will only grow more complex. Antitrust laws must continue to evolve in order to protect competitive markets, which are at the heart of the U.S. economy. If you have any questions or concerns regarding antitrust or merger issues, contact us today at Kessler Topaz. Clients look to us for experienced representation in complex antitrust cases. Courts also frequently appoint our attorneys to act in leadership positions in cases involving anticompetitive behavior.