The impact of AI across many sectors will continue to be undeniable. The power of AI will allow many professionals to work more efficiently, cutting through the noise and giving insights into the important data needed for making decisions faster than ever. However, there are still challenges for AI to overcome including, but not limited to, legal limitations and ethical issues. In this article, we examine the role AI may have in the competition enforcement space and how companies can use AI to ensure they meet their competition challenges.

Antitrust and competition regulators in the United States, the United Kingdom and Europe are expanding the scale and breadth of their investigations, taking a more assertive approach toward overseeing mergers and intervening in competitive practices. Sectors and transactions that previously flew under the regulatory radar are now drawing heightened scrutiny. The demands placed on companies have also broadened, with an emphasis on analyzing substantial data and delivering accurate responses promptly.

Trends in Competition Enforcement

Notable trends seen in the competition space include regulatory bodies taking a tougher stance on noncompete or no-poaching agreements, alleging that they are anti-competitive and can affect market salaries.

The sharing of information between competitors in relation to ESG targets and reporting is another hot topic. The controlling of ESG messaging to the market can influence future regulation on environmental issues and therefore can affect competition.

The use and analysis of data, particularly communication data, which is typically in an unstructured format, is at the center of strategies used by companies to protect themselves against competition issues. That same data will also likely be the most relevant data to a regulatory body, should the company need to disclose on a particular issue. The volumes of unstructured communication data held by many companies is of such a vast scale that it is only by using big data technologies and AI that companies can monitor and analyze these data sets.

AI and Large Language Models (LLMs) can significantly impact the way companies monitor and enforce competition-related best practices; they provide both advantages and challenges. By harnessing AI capabilities, companies can improve data analysis, streamline operations and detect potential violations more efficiently. However, AI deployment raises concerns about privacy, legality, ethics and technological limitations.

Improved Data Analysis With AI and LLM

AI-powered analytics can process vast amounts of data from various sources, helping to identify patterns and anomalies indicative of competition issues, including potential no-poach or wage-fixing agreements. This enhanced capability enables companies to make informed decisions and prioritize resources effectively. Companies can implement AI and LLM directly into their communication systems such as email and chat messages to monitor for potentially anti-competitive language and suggest the use of alternative language to minimize risk, while building an AI-powered monitoring program. What this looks like will depend on the local data privacy regulations and may be sector-specific, but with natural language generation capabilities in LLMs, AI could help facilitate clearer and better structured communication between stakeholders, ensuring that complex concepts are conveyed accurately and coherently.

The LLM can be trained in the latest competition law and in the latest trends seen from the competition regulators, which will give companies real-time analysis of potential competition issues they may face. Advanced AI techniques and deep text comprehension of LLMs enable early identification of potentially problematic conduct, which can prevent further harm and reduce costs associated with extensive investigations.

AI can automate repetitive data tasks, freeing up human resources for higher-level analytical work and decision-making. Enhanced accuracy of data processing will lead to more accurate results and well-informed decisions. Cost savings in data processing and low-level data manipulation are also being seen.

Use of AI in Investigations

The powers of the competition regulators and the activity seen from them over the past few years indicate that even with an AI-powered competition monitoring program, there is the possibility that an in-depth investigation may be needed for a particular issue.

The investigation process may start with a competition regulator sending an information request that can cover a wide range of electronic data sources and an expansive date range. Alternatively, the investigation process may start by the regulator raiding a company in an un-announced inspection. The two scenarios suggest different nuances in how companies should respond, but within both scenarios, the rapid collection and analysis of electronic data to understand potential risks the company may face is paramount.

AI and E-discovery

AI is now integrated into industry leading e-discovery tools for companies and e-discovery experts to deploy in cases. The first iteration of AI within e-discovery tools is designed to accelerate the discovery of key documents and summarize the key findings and facts in the case.

To accelerate review, users of the e-discovery technology can input a description of the matter into the platform. This description can contain any information known at the time, including case background, key individuals and noteworthy terms. The AI then analyzes the document set and provides document summaries for the users to review. The user can then provide feedback to the AI model and improve it's understanding. Across competition investigations, a rapid review of likely relevant material is useful. Deadlines set by regulators are usually aggressive. In a cartel investigation context, applications for leniency have increased over the past two years after a lull. These leniency applications are usually submitted to the regulator within 24 to 48 hours of an announced investigation and, therefore, an understanding of key documents as early as possible is crucial.

Summarizing case findings is another way AI can provide efficiencies within an e-discovery platform. Using generative AI, e-discovery experts can now move directly from document review to crafting case narratives and draft reports containing links and references to the key documents identified. Traditionally these first draft reports would require a read-through of all relevant documents and then drafting from scratch. Now, the AI can give a starting point to verify and work from.

Challenges of AI in Competition Enforcement

The use of AI doesn't come without challenges so companies looking to adopt an AI-driven competition approach should consider the following:

  • Privacy Concerns: Using AI to collect and analyze sensitive company data might raise privacy issues. Strict adherence to data protection regulations and ethical considerations are required.
  • Legal Limitations: AI applications need to operate within legal frameworks, including competition laws and data protection regulations, necessitating continuous monitoring and adaptation to ensure compliance.
  • Ethical Issues: Bias in AI algorithms may lead to unfair treatment of certain companies or sectors, making it essential to implement transparent and accountable processes when deploying AI technologies.
  • Technical Hurdles: Current limitations of the LLM — the size of text they can process, datasets they were trained on, quality degradation in case of large inputs, and insufficiency of computing power to host the LLMs — make the adoption of LLMs slower and costlier.
  • Accountability: Establishing responsibility for AI-driven decisions remains an open question, highlighting the importance of clear governance structures and oversight mechanisms.
  • Continuous Learning: As new forms of collusion emerge, AI systems must continuously learn and update their knowledge base to maintain effectiveness.

Summary

In summary, AI technology has the potential to revolutionize competition enforcement through enhanced data analysis, operational efficiency and early detection of potential violations. However, successfully incorporating AI into competition policy requires overcoming numerous challenges. To maximize the benefits of AI while minimizing its risks, policymakers and organizations must collaborate closely and invest in robust, compliant AI systems that respect regulatory requirements and promote public trust.

As with any piece of technology in the disputes, investigations and regulatory sectors, the value derived from the use of the technology comes from the successful application and interpretation of the results. The power of AI resides in the skilled and experienced human resources that apply and understand the limitations of AI.

Originally published by 28 March, 2024

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.