If one looks around the internet these days, probably even the chatbot ChatGPT [1] itself would be “surprised” to see how many new articles, essays, blogposts, personal posts and videos discuss a plethora of issues surrounding the use of artificial intelligence (AI), and specifically, generative AI.

This blogpost does not aim to fully explore the various legal aspects of AI, but rather wishes to focus on one narrow intersection: namely, the one between generative AI tools and competition law. Below, we will therefore discuss – or rather: present some initial questions – as to how each main area of competition law could be relevant to the business conduct of ChatGPT (or rather: OpenAI Inc., the company behind the chatbot) and similar generative AI tools. [2]

Anti-competitive agreements

The prohibition of anti-competitive agreements (or in other words: the cartel prohibition) clearly applies to all undertakings: this naturally includes companies engaged in providing generative AI services, such as ChatGPT itself. However innovative these services are, it is obvious that the companies involved (including their managers, employees, etc.) are clearly prohibited from engaging in any collusive conduct (eg concluding agreements on the fixing of prices, sharing of markets or exchanging sensitive information).

Now, are there any specialities in the generative AI business in this respect? The following points already appear to merit special attention:

  • generative collusion: an extremely interesting question is the way and extent competing generative AI tools can “use” each other, learn from each other and thereby engage in some kind of “interaction”. For example, if one chatbot can access the public business conditions (including prices and terms) applicable for a competing chatbot, it can then suggest (or possibly: make) changes to its own business conditions. This, in itself, can be regarded as autonomous adaptation to market conditions, which should not raise concerns. If, however, this is practiced by most or all chatbots on the market, could the chatbots realise the advantages of cooperation as opposed to competition and thus achieve a homogenity of general terms and conditions or prices?
  • platform collusion: A chatbot could provide a platform for collusion for players in an entirely different market, especially if it can collect and disseminate current data. [3] For example, competitors in an industry could provide the chatbot with their current pricing, output and/or other sensitive data, and then acquire data about others by putting relevant questions to the chatbot. In such an instance, does the chatbot bear liability – just like once a well-known consultancy firm had to bear for organising secret physical meetings for the worldwide hydrogen peroxide cartel – as a contributor under EU competition law? [4]
  • cross-shareholdings / cross-directorships: it appears that various Big Tech companies and other investors hold shares in various generative AI companies (for example, Amazon Web Services was one of the founders of OpenAI, while it is also engaged in operating Hugging Face, a platform for developers to collaborate in generative AI systems). As a result, these companies – who may be direct competitors to each other – need to pay particular attention to erecting internal and external barriers to prevent the exchange of sensitive information (eg business plans, current / future pricing data, etc). These compliance measures appear to be essential to ensure to enable fully autonomous decision-making at each generative AI company and thus to secure a fully competitive landscape.
  • compliance: ChatGPT could also be an incredibly useful tool to foster competition law compliance, for example, by generating or reviewing simpler compliance documents for companies (eg drafting a new manual to handle „dawn raids” by competition authorities). Certainly, this has to be closely supervised by a trusted competition law expert to ensure accuracy and completeness. In fact, using AI for compliance is not even a new phenomenon: DLA Piper has already been using its own AI tool, AIscension to quickly spot evidence of anti-competitive behavior in large batches of documents (such as entire email boxes).

Abuse of dominance

Competition law also prohibits abuses of a dominant position. Could ChatGPT or other similar service providers be regarded as dominant, either individually or collectively?

On the one hand, ChatGPT is an extremely popular tool with a reported more than 100 million users in February 2023 [5], while on the other hand, there are also major competitors to OpenAI’s new system, such as Google Bard, Facebook’s RoBERTa or Amazon Web Services’ Hugging Face and also innovative start-ups like Caldera Labs Inc.’s Cohere or AI21Labs’ WordTune.

  • First, to be able to provide a bit more nuanced view on possible abuse issues, we would need to review the precise arena of competition for generative AI services, ie the relevant product and geographic market. To this question, ChatGPT itself provided the following response:

 In the above excerpt, ChatGPT – although with clever disclaimers… [6] –appears to correctly position itself as a provider of digital services, putting itself on the narrower market for natural language processing and generation. At the same time, it has been argued that ChatGPT could entail disruptive competition on a wider scale and may very well put competitive pressure on the market of general internet search (such as Google Search) as, from the perspective of the consumers, it is substitutable – and may be even superior – to such general search services. [7]

  • Second, it is to be noted that although the market for natural language processing and generation is indeed innovative, there appear to be significant barriers to entry, which include, for example, the exponentially increasing costs of training the large language models (LLMs) that are used by generative AI companies. [8]

  • Third, a possible indication of dominance could be that generative AI services may also fall under a special regulation, namely Regulation 2022/1925/EU, or – as it is more widely known, the Digital Markets Act. It is even possible that companies providing generative AI services may be designated by the European Commission as so-called “gatekeepers” as they could be argued to provide so-called “core platform services” (such as “virtual assistants” and “search engines”).

In light of the foregoing, the question of possible dominance is certainly too early to call. However, the relevant market players will have to be watched closely to ensure that they indeed remain open, innovative and competitive. [9]

Merger control

There is already significant investment by major tech players into various generative AI products, including ChatGPT. Certainly, if these investments reach the level of control as defined in most merger control regimes (eg majority of voting rights or a veto right on important business decisions), then these transactions (as concentrations) could be subject of scrutiny by antitrust authorities.

So far, we are not aware of any transactions notified to a major competition authority, where generative AI would have been at the forefront of the competition assessment. It is, however, expected that sooner or later merger control will also play a role in this field of the economy as well. In such a case, the above-mentioned questions of the relevant market, and the possible concentration of the generative AI sector will likely be highly important.

Unfair competition

A number of domestic laws (eg Germany, Hungary) include various overall prohibitions on unfair competition, outside the scope of the classic antitrust world. For example, in Hungary, the Hungarian Competition Act prohibits unfair competition in general, and then also enumerates further specific prohibitions such as the one relating to passing off (preventing one trader from misrepresenting goods or services as being the goods and services of another (competing) trader) and the one relating to the breach of business secrets.

  • First, one issue with generative AI is that it is extremely easy to produce text – intentionally [10] or unintentionally – that resembles the style of other companies/brands. In this case, ChatGPT (or, even more plausibly, ChatGPT’s sister product, DALL-E the image generator) can be regarded as a direct competitor to those artists or designers whose style it is capable of mimicking. As a result, when ChatGPT produces a new text in the distinctive style of an artist, and if such text is then used by a company (eg as a slogan for a new commercial), such scenario could entail harm to the artist involved (and thus a significant loss of business). This scenario could therefore be described as unfair competition to the artist in the form of passing off. Importantly, this prohibition could also apply even if, from an IP perspective, there is no copyright infringement (eg due to the fact that under copyright law, as opposed to specific works of art, a mere style / stylistic elements are much more difficult to protect).

  • A second issue is the breach of business secrets. Under EU law, trade secrets are defined by Directive 2016/943/EU as information that is secret, has commercial value because it is secret, and it has been subject to reasonable steps under the circumstances, by the person lawfully in control of the information, to keep it secret.

  • EU law also specifically prohibits the unlawful acquisition, use and disclosure of such trade secrets. What if ChatGPT learned of a trade secret during its training (eg due to information made inadvertently available on the internet)? What if ChatGPT learns of a trade secret as a result of a new data being inputted into it (eg if a lawyer copy-pastes a client’s long contract to have it reviewed for specific clauses)? [11] What happens if ChatGPT then reveals such newly learned information / secrets to an interested third party? Arguably, such an “action” by ChatGPT could entail a breach of the above prohibitions and could thus constitute unfair competition.