Author: Ashima Sarin, a student of Maharaja Surajmal Institute, GGSIPU
ABSTRACT
Artificial intelligence (AI) has quickly emerged from being just a piece of technology into an important factor influencing the economic activity of enterprises. Given the increased role of AI, there are issues connected to market power, access to data and computing power and the potential power of large firms to manipulate competition. Due to the concentration of critical infrastructure of AI among a few technology companies, the authorities of the whole world are now considering if existing competition legislation is adequate enough to tackle new problems.
The present article investigates the connection between competition law and AI markets, taking into account the issues of data aggregation, market entry barriers, self preferencing and algorithmic coordination. Moreover, it assesses if current antitrust regulation, especially in digital markets, is enough to regulate competition influenced by AI. The main idea of this work is that even if competition laws are still adequate in most cases to regulate anti-competitive behaviour on the markets of AI technologies, they need some changes.
TO THE POINT
Who controls the future of artificial intelligence?
The solution to the problem is no longer solely an engineering and technology question but also one for competition law. For modern AI systems to work, vast amounts of data, advanced computing infrastructure, specialized manpower and financial resources are necessary. This would obviously favour big tech firms with sufficient resources.
Since AI is being implemented in search engines, cloud computing, online advertising, e-commerce and productivity software, it has become a question whether dominant firms can utilize the existing market power in order to solidify their position in AI markets. Thus, the main question arises about whether the existing antitrust regime will be able to preserve competition in the industry moving faster than most legal systems.
USE OF LEGAL JARGON
Competition law assesses the market behaviour through such notions as “relevant market,” “dominant position” and “anti-competitive agreements.” “Abuse of dominant position” means actions by a dominant company harmful for competition. “Market foreclosure” means that competitors are hindered from properly competing in the market. “Network effects” mean that the more users of a certain service, the more valuable the service becomes. “Self preferencing” means preference of the company’s products over those of competitors on the platform.
THE PROOF
AI Markets and the Creation of Novel Forms of Market Power
Artificial intelligence has evolved from being a specialized field of research to a commercially viable form of technology that may affect whole sectors. Nowadays, AI tools help to write, program, interact with customers, conduct health care research and support businesses in making decisions. Nonetheless, creation of more complicated AI tools requires certain resources that are concentrated in the hands of just a few companies.
Unlike some conventional markets, development of AI depends very much on access to several key factors, namely data, computing capabilities and funds. Development of sophisticated AI requires large amounts of data and high levels of computing power, which are often offered by big cloud computing. Therefore, companies that have access to all those factors have a great advantage in terms of competition.
However, such a situation creates a problem of market power. Competition law does not forbid success and innovation. Yet it aims to avoid a situation in which dominant companies abuse their dominance and restrict competition in any way. Concerning the AI market, the issue is not about whether large companies should succeed but whether control over crucial resources makes competition impossible.
Data Aggregation and Barriers to Entry
Data is said to be the oil of the digital economy. The more data an AI system gets, the better its performance becomes. This is why companies with strong user bases and digital environments tend to accumulate more data compared to their rivals.
This poses a serious problem for new businesses. Innovating start-ups might find themselves lacking data to match the capabilities of incumbents in the industry. That is why competition regulators have turned their focus on data accumulation as a market power tool. Data ownership per se cannot be anti-competitive, but when there is too much of it and it serves as a barrier to entry, problems appear.
Computational Resources and Market Access
Apart from the raw data, advanced AI technology also needs a lot of computing power. Large scale AI model training requires specific hardware and access to the cloud computing infrastructure that will be able to process huge amounts of data.
The clustering of cloud computing infrastructure within a few technology companies has raised antitrust issues. Where control over the necessary infrastructure rests in the hands of a few companies, the smaller companies can find themselves in a situation where they will depend on those companies for the provision of services.
This can lead to issues of market foreclosure and lack of competition. Antitrust bodies increasingly look into how control over the infrastructure can allow companies to use it to gain an advantage over other companies through competition.
Self-Preference and Algorithmic Issues
One of the major competition issues in digital markets is self-preference. Here, the dominant platform favours its product or service over others, putting the competitors at a disadvantage.
In AI markets, self-preference can emerge where companies incorporate AI based solutions into existing digital ecosystems and prioritize such solutions over rival solutions. These actions affect consumer preferences and make it more difficult for competitors to compete.
The next important issue is algorithmic coordination. Traditional competition law deals with overt agreements between competitors. AI based systems, on the other hand, are capable of analyzing market dynamics and changing strategies accordingly. While purely autonomous algorithmic collusion does not seem to exist yet, regulators are aware of possible competition law issues associated with more advanced algorithms.
Is Regulation of AI Markets Possible Within the Present Antitrust Regime?
An often made claim is that AI is so innovative that completely new competition laws need to be enacted. Such an assertion fails to note an important aspect about the antitrust regulation system: that of being technology neutral in nature.
Competition law is not concerned with regulating technologies per se but rather anti competition itself. Irrespective of whether the monopoly power of a company is built through the railways, telecommunications, internet or AI, the basic issue remains whether the process of competition itself has been hindered.
In India, the Competition Act, 2002 offers the possibility of regulating any anti competitive agreements, abuse of dominant position, and mergers. It has enough wide provisions that can cover conduct related to AI. In case of any anti competition behaviour by a dominant company working in AI technology, such a regulation can be addressed using current provisions of the act.
The same approach is being followed in Europe and America as well. New antitrust regimes are being created but in the context of new technology, existing laws of competition have been used.
Challenges for Competition Agencies
Though existing legislation continues to apply, there are some difficulties involved in using it in relation to AI markets.
First, it is quite challenging to delineate the scope of a particular market. AI solutions usually serve several purposes at once and operate in interrelated sectors, thus making delineation increasingly difficult.
Second, innovations tend to outpace the process of regulation. Investigations related to competition law enforcement are likely to take years, whereas AI markets might change significantly in just a few months.
Third, due to the high level of technology, it may be hard for the regulator to evaluate competitive impacts since special expertise in dealing with the issues is required.
None of these challenges implies that existing laws lack validity. Instead, this points to the necessity for competition agencies to come up with new approaches based on the laws.
CASE LAWS
1. Competition Commission of India v. Google LLC
Competition Commission of India opined that the activities of Google in the context of the Android ecosystem indicated potential abuse of dominant position. The case showed the connection between the control over the digital ecosystem and the competitive process and also showed how Indian competition laws could handle the conduct within the technology based markets.
2. Google Shopping Case
The Google Shopping case is an important decision where the European competition authorities decided that Google had abused its dominant position through its practice of self preferencing by showing preference to its own comparison shopping service and disadvantaged the competitor’s service. This case became an important precedent for the notion of self preferencing in modern AI markets.
3. United States v. Microsoft Corp.
The Microsoft case remains one of the most significant technology antitrust cases. It involved a dispute related to the claim that Microsoft was using its market dominance in the operating systems to dominate the browser market.
CONCLUSION
The development of artificial intelligence is transforming the market and industry landscape in a rapid manner. Although artificial intelligence provides ample opportunities for innovation and economic growth, there are certain issues pertaining to market dominance, barriers to entry and concentration of economic power.
Traditional antitrust policies are generally well equipped to regulate markets in artificial intelligence since they rely more on conduct than on technology. The concepts of abuse of dominant position, market exclusion, anti competitive agreements, and consumer welfare continue to serve as powerful instruments for dealing with competition issues. Nonetheless, the effectiveness of competition law in the age of artificial intelligence would largely be dependent on the application of these concepts in highly complex digital environments.
FAQ
Q1. What is competition law?
Competition law refers to an area of law aimed at ensuring healthy competition and prohibiting activities that limit market competition or negatively affect consumers.
Q2. Why is AI a cause for competition issues?
The development of AI technologies usually relies on the availability of certain data, computing power and capital that could become highly concentrated in a limited number of companies.
Q3. What is self preferencing?
Self preferencing means that a dominant company favours its products and services over other companies’ services.
Q4. Could existing competition laws apply to AI companies?
Yes. The current competition laws are largely technology neutral and could deal with anti competition activities independently from the type of market.
Q5. Is there a need for completely new antitrust laws?
Not always. All competition issues related to AI markets could be resolved by the existing legal framework.
