In the light of an eight months study into the AI foundation models and the publication of the first report in October 2023, the UK Competition and Markets Authority (CMA) has offered an updated report alongside a technical report, emphasizing the fact that AI competition can’t work is not what it is supposed, the way it should. In a speech attended on the 11th of April 2024, the CMA CEO Sarah Cardell explained that the CMA had crafted a program and started work in that direction.
The CMA’s findings
Hereinafter, we will present the core issues and key requirements that CMA claims AI companies should consider as principles while developing their business practices. Therefore, it is plausible that the CMA will increase its enforcement activities in this area, unleashing all the arsenal tools of state regulation—combining actions on mergers, competition, and consumer enforcement with market investigations and digital control. We anticipate the CMA’s report will be the prelude to further US sanctions from the European Union (EU) and US agencies.
The CMA that the foundation AI chain is related to the digital firms in its report employs a web of partnerships and many acquisitions such as acquisitions, where the application is not the main aim but acquiring the expertise of other firms. As a result, the C CMA claimed that these business collaborations and investments drive firms to deepen their relationships in the AI supply chain, and they might involve financial support or provision of some key input requirements like data and compute, as well as exclusion or priority distribution schemes.
The CMA is worried that these partnerships would allow big digital players to take onto themselves all the key parts in the value chain of AI and further strengthen the market position, which the CMA already sees as strong (or extend this influence onto adjacent markets). In addition, the updated report considers the chances that players in the market might try to shape such collaborations to suppress their competition. In this context, the CMA has noted that it could be particularly concerned if a partnership has one or more of the following features: In this context, the CMA has noted that it could be particularly concerned if a partnership has one or more of the following features:
Whereas the most popular party presides over an important input source in AI development, or where the parties might decide to install the models, the other party may have an existing power downward. Details associated with featured models can form a strong decision-making factor in the long term.
In the case of the models, the joint engagement of the parties results in either party’s control of the models and/or deployment of the other’s technologies and perhaps emerging opportunities for embracing incentives.
Raising its worry that several investments have been shaped as lettings to buy, thus violating the merger control rules, the CMA has informed the community to be more vigilant in using its merger control powers to examine whether such partnerships (and, more so, acqui-hires are ‘mergers’) and are harmful for the competition. The other institutions will surely be ready to run reviews that might take other means away from the current use.
However, while the CMA has moved closer to the North American stance, the core statement considers only AI-based investments and skimps on issues outside that particular investment scope. Although passing as an office to the investment ecosystem, the CMA fails to delve into the pros and cons they seem to provide, leaving a rather shallow analysis of this correlation. From the study of CMA, it is evident that several of these partnerships and parties are also quite well known. Meanwhile, it will be up to them to closely examine these two articles of association and the companies involved to forestall this lapse.
From the general context, it ought to be the CMA monitoring team that will be 100% intensified behind the acquisitions and the major injections in the sector that have not been notified. Investors and deals in the AI space may face CMA scrutiny over their investments and acquisitions, even where those investments may initially seem limited, or the market does not meet the normal merger thresholds.
Way forward for CMA
The CMA can, therefore, step in even in investments that are not considered of national significance. British merger control threshold is satisfied by utilizing merely the minority stake that allows the said controller to exercise ‘material influence,’ which is not necessarily capital contribution alone but also the representation rights of the board and commercial or financial arrangements between companies.
Other jurisdictions have comparable institutions for considering the cases where the thresholds above seem left out. When merger control powers are inapplicable, the CMA and other regulators will likely examine partnership-type arrangements within the normal competition law framework, which IPPC permissibility decisions will not replace.
So, even though investments into the AI sector are anticipated to grow in volume, companies seeking funding or planning a strategic investment should conduct an early regulatory risk assessment and consider whether the specific characteristics of the deal are likely to be especially attractive for the regulators from European jurisdictions (e.g., EU, Germany, Austria or the UK) being happy with demanding only simplified requirements for the pooling of various national units. The assessment undertaken should determine the risk-sharing clauses in the contract documentation. To accomplish this analysis, it is necessary to point out the aims of such partnership or transaction; if this comes into question before the appropriate regulator, there is no choice but to highlight that competition-promoting effects are articulated.
Moreover, with the emerging substance trends and the fact that both the CMA and other regulators will take into account the historical ‘horizontal’ distance of the transaction, which is measured in the contracts’ terms, the long-term time horizon of the evaluations up to several years in the future will go beyond the mere decision-making process frame.
Such a review is often composed of a thorough review of internal documentation, especially deal rationale or another party’s strategy in this business field. It also considers the CMA’s latest concern, arising from the fact that many large digital firms are monopolizing the fund of critical data inputs.
