You’ve probably all seen the terror ChatGPT has unleashed on white collars, from research analysts to copywriters, over the past two months. What about lawyers, economists, or paralegals working in competition law? Let’s summarise for those who haven’t heard yet: ChatGPT is a chatbot created by OpenAI. It became the fastest-growing app in history, reaching one million users in five days and 100 million in two months. When asked to make a moral judgement, a political judgement, or a numerical calculation, ChatGPT begins its response by reminding us of what it is: ChatGPT is a language model! We should adjust our expectations accordingly. Although it has not yet started to eliminate professions, it provides incredible help to those working in many fields. This is what I want to discuss in this blog piece: What should competition lawyers and economists expect from ChatGPT?
Let me warn you about ChatGPT’s responses right away:
- It does not owe us the truth. It may base its evaluation on incorrect information. OpenAI reminds you of this every time before using ChatGPT.
- It does not yet claim to be able to make calculations.
- There is a reported bias problem in its answers.
With this in mind, I asked ChatGPT how it can assist a competition lawyer in a merger control proceeding. The answer went as follows:
– “A competition lawyer can use OpenAI’s services in a variety of ways during a merger filing to a competition authority. Some potential uses include:
- Research assistance: OpenAI’s natural language processing and machine learning capabilities can be used to gather information and data on relevant market conditions, industry trends, and regulatory developments.
- Document review: OpenAI’s document analysis capabilities can be used to review large volumes of documents related to the merger, including financial reports, contracts, and legal filings.
- Predictive analysis: OpenAI’s predictive analytics capabilities can be used to analyze the potential effects of the merger on competition, such as market concentration, pricing dynamics, and innovation.
- Communication support: OpenAI’s language generation capabilities can be used to support the drafting of legal arguments and submissions to the competition authority.”
When I asked how he could help in cartel cases, it added the following point:
“Data analysis: OpenAI’s machine learning and natural language processing capabilities can be used to analyze large volumes of data, such as emails, contracts, and financial records, to identify potential evidence of collusion or anti-competitive behaviour.”
And it concluded with a disclaimer for both questions:
“It’s important to note that while OpenAI’s services can provide valuable support during a merger filing, the final decision on the approval or rejection of the merger will ultimately be made by the competition authority, taking into account a range of legal, economic, and policy considerations. Additionally, competition law firms may have to be mindful of ethical considerations when using AI in the legal context, such as data privacy, bias, and accountability.
It’s also important to note that while OpenAI’s services can be useful tools for competition lawyers, they are not a substitute for professional expertise and judgment. Competition lawyers should still use their own expertise and discretion when evaluating the relevance and reliability of the information generated by OpenAI’s services, and in making strategic and tactical decisions related to the merger filing.”
That is refreshing! So, our jobs are safe for now.
ChatGPT Assists Hypothetical Merger Case
Since I had already been warned that ChatGPT’s answers were unnecessarily verbose and sometimes overly ambitious, I tested it by creating a hypothetical merger case to see if it could actually do what it said. I asked the following questions:
“VW AG and Daimler AG are about to merge, and I am going to notify the merger to EU Commission. Could you please prepare a note answering the following questions? What are the relevant product markets that I will address in the notification? What will be the merged company’s total market share in these relevant markets? What competition concerns might arise? Will they be predominantly single-firm dominance issues or more of coordinated effects? Could you also do research on the jurisdictions that we need to notify?”
– “Sure, I can help you with that!
Relevant product markets: The relevant product markets will depend on the specific products and services that are being offered by the merging companies. However, based on the information provided, it is likely that the relevant product markets will include:
- Passenger cars: This market includes all types of passenger cars, such as small cars, mid-size cars, and luxury cars.
- Commercial vehicles: This market includes all types of commercial vehicles, such as vans, trucks, and buses.
- Automotive parts and accessories: This market includes all types of automotive parts and accessories, such as engines, transmissions, brakes, and electronics.
Merged company’s total market share: The market share of the merged company will depend on the specific markets and products involved. However, VW AG and Daimler AG are both major players in the automotive industry, so it is likely that the merged company’s market share will be significant in many markets.
Competition concerns: The competition concerns that might arise from the merger will also depend on the specific markets and products involved. However, the European Commission will likely be concerned about the potential for the merged company to gain significant market power, which could result in reduced competition and higher prices for consumers. This could be a single-firm dominance issue or a coordinated effects issue, depending on the specific circumstances.
Jurisdictions to notify: The merging companies will need to notify the European Commission, as well as the competition authorities in any other jurisdictions where they have a significant presence. This will likely include the competition authorities in the countries where the merging companies are based (Germany, in this case), as well as any other countries where they have a significant market share. It may also be necessary to notify the competition authorities in other regions, such as the United States or China, if the merged company will have a significant presence in those markets.”
ChatGPT Disrupts Economic Consultancy
Then I wanted to know if competition economists like me should be scared: “What do you recommend to the competition economists working on this merger?”
– “As an AI language model, I am not able to provide specific recommendations to the competition economists you hired for this case, as I do not have access to the specific details of the merger or the results of the competition economists’ analysis. However, I can provide some general guidance on what competition economists might consider when analyzing a merger such as the one between Volkswagen AG (VW) and Daimler AG:
- Identify the relevant product and geographic markets: Competition economists should first identify the relevant product and geographic markets in which the merging firms operate. This will help to assess the level of competition in these markets and to determine whether the merged company would have significant market power.
- Assess market concentration: Competition economists should assess the level of market concentration in the relevant markets, using appropriate measures such as the Herfindahl-Hirschman Index (HHI) or market shares. This will help to determine the degree to which the merger would increase concentration and potentially reduce competition.
- Analyze entry barriers: Competition economists should analyze the potential for new firms to enter the relevant markets, and whether the merger would create barriers to entry that could prevent or limit competition.
- Evaluate potential anticompetitive effects: Competition economists should evaluate the potential for the merger to have anticompetitive effects, such as higher prices, reduced quality, or decreased innovation. This analysis should consider the specific details of the merger and the market conditions in which the merged company would operate.
- Consider potential efficiencies: Competition economists should also consider the potential efficiencies that could arise from the merger, such as cost savings or increased innovation, which could benefit consumers.
Overall, competition economists should take a thorough and balanced approach to analyzing the merger, considering both the potential anticompetitive effects and the potential efficiencies. They should also be prepared to present their analysis clearly and persuasively to the European Commission or other competition authorities, if required.”
Conclusion
ChatGPT sounded to me like an ambitious and enthusiastic employee with good ideas, but still a long way to go. It is far from being able to provide the services it claims, at least for now. However, it is still in the testing phase and already disrupting many industries and professions. When its creators develop specialised AI products for service firms such as law firms, it will certainly reduce, if not completely eliminate, the number of human employees. AI services, which will tremendously increase the productivity of the remaining employees, will make it mandatory for law firms and consultancy companies to change the scope of the services they provide and the required qualifications of their new employees.
At this point, the main issues that haunt my mind are as follows:
- Confidentiality: It goes without saying how important confidentiality is in our work. Every time you start using ChatGPT and entering and analysing your client’s data, you share many issues, each of which constitutes a trade secret, with a private company.
- Verification: Let’s imagine that ChatGPT progresses further, which I don’t think will take very long, and can decide which analysis it should do for which situation. This analysis could be a legal analysis or a quantitative analysis. Even in this scenario, someone will need to check ChatGPT’s methodology, its calculations, the integrity of the data it uses and, consequently, its analysis. I don’t think we could say to a public authority or a court that “the fault lies with our AI provider, not us”.
We shall not flag nor fail. We shall go on to the end. We shall never surrender!
…
Originally published at Kluwer Competition Law Blog
