Some recent mergers, acquisitions and investments in the business world have highlighted the strategic value of data to companies. These businesses are not just buying assets or market share—they are also acquiring or investing in large, complementary datasets. This process is known in the business world as horizontal integration.

This integration can drive innovation and provide competitive advantages. It can also open up new revenue streams. Some examples include Microsoft's acquisitions of LinkedIn and GitHub as well as Amazon's acquisitions of WholeFoods and the Washington Post. Then there has been Discovery Communications' merger with Warner Brothers, IBM's investment in Hugging Face and Google's investment in Anthropic.

As the last two examples illustrate, data is extremely important for AI companies. It's vital for "training," or improving, AI systems. Training AI systems on large, new, varied data sets allows companies to develop more advanced, more powerful AI systems.

But against the background of this scramble, there is also a growing consensus that some form of regulation is needed to address the ethical, safety and fairness concerns associated with AI.

But regulating AI presents a unique set of challenges. This is mainly due to its foundation on intangible elements such as software and algorithms. These elements can be easily modified, replicated and distributed across borders with few physical traces. This helps them evade traditional regulatory mechanisms that rely on controlling physical goods or specific locations.

Yet a promising approach to regulating AI is one that would focus on controlling access to the very data that is the lifeblood of AI development. Since data is behind the rise of horizontal integration as well as fueling the growth and sophistication of AI systems, its concentration in the hands of a few entities can lead to monopolistic dominance. In short, it gives too much power to too few companies.

Antitrust model

To mitigate this, regulatory frameworks could be designed that resemble existing antitrust laws—but focused around data aggregation. They would help ensure a diverse and competitive landscape in the access to data. By preventing any single company from amassing an overwhelming data advantage, these regulations would aim to foster a more balanced field. Innovation must be allowed to thrive without being stifled by monopolistic control.

To properly achieve this outcome, we suggest that regulators need to look at limiting horizontal integration. As AI technologies continue to evolve and the demand for diverse and extensive datasets grows, companies will increasingly be motivated to pursue horizontal integration.

This trend towards integration not only consolidates data assets but also potentially reduces competition, as fewer companies come to control larger shares of valuable data. Therefore, regulatory scrutiny of such mergers and acquisitions becomes essential to ensure a competitive landscape where data does not become excessively concentrated in a few hands.

It's important to note that the trend towards horizontal integration is already moderated to some extent by regulatory and ethical considerations, particularly around data privacy and existing antitrust laws. These considerations play a critical role in shaping the extent and nature of integration.

This article is republished from The Conversation under a Creative Commons license. Read the original article.