From Innovation to Domination: The Anti-Competitive Edge of Ai

In recent years, the world witnessed an exponential surge in the use of Artificial Intelligence (AI) and its use in almost every realm of business. AI has been a key asset in rapid expansion and technological development of various sectors and industries. The disruptive potential of AI has unlocked aspects which were previously considered as unreachable or out of human realm.

In this article, the author endeavors to shed light on the nexus between AI and Competition Laws and the potential anti-competitive risks by use of AI. From the current competition standpoint, it might be difficult to exactly determine the anti-competitive threats that loom around the usage of AI, given the complexity & technicality involved.

Possible anti-competitive risks associated with AI use

Following are some of the possible anti-competitive risks associated with AI –

1. ALGORITHMIC COLLUSION –

Particularly in the digital markets, there’s a risk that companies may deploy certain such algorithms in their software or back end of the websites which will be pre-programmed to modify the price of product/services automatically based on any kind of changes in the market. If every competitor uses such a system, there are very high chances of price parallelism since every competitor’s system will adjust itself to each other’s price and market movements on a real time basis. This will facilitate tacit collusion between competitors i.e. there will not be any direct link or communication among competitors themselves, just passively coordinated acts without even a single contact between them. The algorithm will automatically adjust the changes based on each other’s behavior. Such coordination among different algorithms will prove to be fatal to fair and equitable competition.

1. MARKET FORECLOSURE / EXCLUSIONARY PRACTICES

Exclusionary practices refer to the conduct when an entity prevents other market players from entering the market or forecloses the market for competitors to capture a greater market share.

The generative AI may be used by a dominant entity to prevent other competitors (including small competitors) from entering that sector/industry/market to which the dominant entity belongs. The algorithm may be coded in such a manner that by default it gives preference to the select competitors which are in any way related to the dominant entity. These select competitors may include partners, brand sponsors, subsidiary company or sister’s concern entity.

In the competition arena, one of the well-known exclusionary practices is search engine manipulation. Google has faced various antitrust penalties across the globe for manipulating search engine algorithms to favor its own product over competitors, which leads to limiting market access for competitors. Exclusionary practices by Google also include AI driven targeted ads which prevent competitors from reaching potential customers. This leads to greater market dominance.

1. NETWORK EFFECTS AND BARRIERS TO ENTRY

As mentioned in point (a) above, the AI collects, processes and uses the data it gathers for further optimizing its performance and output. As a result, the entity owning the AI ends up having a vast amount of data which can be used to predict market trends and manipulate the market in its own favor. The said entity gains an edge over the competitors due to the data it possesses. Having such data may unleash significant patterns that are not available to the other small competitors. As a result, these smaller competitors find it tough to compete with this big dominant entity. This leads to a barrier to entry for these smaller counterparts since they do not match the data sophistication nor the tools which the big dominant entity has.

PRICE DISCRIMINATION

Some entities may use AI for preparing price algorithms. Price algorithms can be understood as a digital tool which analyzes various factors to determine the price of the product. Typically, these pricing algorithms are based on thresholds set by the entity. The factors for price determination include market type, product type, category of customer, customer preferences, market share of the entity, among other such relevant factors.

The price algorithms will enable the entity to calculate prices in such a manner that might be unfair to customers or other competitors. Typically, there are three types of price discrimination which an algorithm can perform –

1. Dynamic Pricing – It is based on real–time updates in the market. The category of updates could range from demand of product, supply of product, consumer preference, competitor status etc.

1. Personalized pricing – In this, the price is calculated differently for different types of consumers. Each consumer sees a different price for the same exact product on the same platform. The prices to be shown to consumers are totally dependent upon consumer behavior, purchase history, location etc.

• Algorithmic pricing – In this type of pricing, the price is estimated via the pre-programmed code embedded in the back end of the website hosting the platform.

AI under the radar of Competition Authorities

Currently, AI is getting significant traction from the competition authorities since they are delving deep into the AI space to fully understand the potential antitrust risks associated with AI on market competition.

This is evident from the following steps taken by various Competition Authorities around the world –

1. In May 2023, the UK Competition Authority, Competition and Markets Authority (CMA) has conducted a preliminary review of AI Foundation Models, the main purpose of which was to assess the risks associated with such type of models from the competition law lens. Basis the review, the preliminary report was published.
2. In European Union (EU) region, the European Commission (EC) invited contributions on competition in virtual worlds and generative AI and sought information from various stakeholders on this issue.
3. In France, the Autorité de la concurrence (French Competition Authority) started enquiry and launched a public consultation on February 2024.
4. Recently, a Joint Statement was issued by EC, CMA, Dept. of Justice US and FTC on Competition in Generative AI Foundation Models and AI Products. This reflects the commitment shared by the authorities to address this issue.
5. On the similar lines, our Indian Competition Regulator, CCI, recently announced that it is launching a market study on AI and Competition. The Ministry of Corporate Affairs (MCA), Government of India established a Committee on Digital Competition Law (CDCL) which published a report citing its recommendation on this aspect. As a result the draft Digital Competition Bill 2024 was introduced along with the said report.

Possible Mitigation Strategies

Following are the strategies to be adopted to curb the anti-competitive effects of AI –

1. ALGORITHMIC AUDITS –

Conducting audits of the algorithms and the whole AI systems at regular intervals will enable authorities to keep a check on non-compliance issues and anti-competitive effects by the algorithm. The legislation should provide for mandatory checks to be performed and for a report to be submitted to the relevant competition authority.

1. TRANSPARENT FRAMEWORK

The entities using AI systems should bring transparency in their AI systems’ operation. There should be a dedicated page on the entity’s website describing the mechanism of its system and how its algorithm functions. All the relevant details about the working pattern of the algorithm should be available on the entity’s website.

1. AI FOR ANTITRUST ENFORCEMENT

Authorities should use AI itself for curbing the anti-competitive effects emerging from the use of AI systems. Using AI for this purpose will definitely give an edge to Competition Authorities in maintaining the market equilibrium and avoiding concentration of power in the marketplace.

1. STRICTER PRIVACY LAWS

Laws pertaining to data privacy/data protection should be strengthened to enhance the grip of authorities on potential data leaks/data manipulation leading to anti-competitive practices. One thing that can be implemented is data anonymity of personal data to prevent the abuse of consumer data in anti-competitive practices. Also we can create a free flow channel between various entities to reduce the dominancy and dependency of one single entity. These channels will facilitate the transfer of data from one entity to another. The consumer will be free to switch to another platform without any hassles. This will prevent an entity in becoming the monopoly.

1. REG TECH

Reg Tech (short for Regulatory Technology) refers to the use of technology in regulatory and compliance related activities and to leverage AI in compliance tasks. Such AI analysis systems should be created by authorities to detect any symptoms of potential anti-competitive threats before they happen. The algorithms should be coded and designed in such a manner that it will enhance the ex-ante capability of authorities. Blockchain technology can also be used for transparency in the competition domain.

AI has been growing at an exponential pace. The Competition Authorities are relentlessly trying to understand the implications of AI from the competition perspective. Although AI does have a plethora of pro – competitive effects, the way it functions, the likelihood of creating anti-competitive effects using AI is very high. It will be too early to say about the exact anti-competitive effects of AI, nevertheless this article lays down a synopsis of the possible anti-competitive effects of using AI. In the Indian context, The Digital Competition Bill 2024 is brought by CCI to tackle the new age AI and digital issues in the competition arena. Further, the best way to tackle this issue is to use AI itself as a remedy to counter the possible anti-competitive practices. AI if used properly, will enhance the capability of CCI in enforcing and tackling competition issues in digital space.

Author: Ashutosh Shukla, in case of any queries please contact/write back to us via email to [email protected] or at IIPRD.