CMA raises red flag over competition concerns in AI foundation models
CEO expressed worries over Big Tech's capacity and motivation to skew Foundation Model markets
The UK's Competition and Markets Authority (CMA) has raised concerns about competition in the AI Foundation Models (FMs) market.
During a speech in Washington DC, CMA CEO Sarah Cardell highlighted the challenges arising from the growing dominance of a few major Big Tech companies and their potential to manipulate the market to their advantage.
In a recent Update Paper on foundational AI models, Cardell expressed serious reservations about the sector's progress.
Speaking at the conference, Cardell highlighted the transformative potential of FMs, describing them as a possible "paradigm shift" for societies and economies.
However, she also pointed out the increasing dominance of a select group of tech giants, including Google, Apple, Microsoft, Meta, Amazon (aka GAMMA) and Nvidia in this field.
These companies hold significant positions in both FM development, supplying essential components like computing power, data, and skilled personnel, and in FM deployment, controlling crucial access points such as apps and platforms.
Cardell expressed worries that these firms have both the capacity and the motivation to steer these markets to their advantage.
Such influence could severely undermine fair, transparent, and efficient competition in FM-related markets, potentially disadvantaging businesses and consumers by limiting choices, diminishing quality, and raising prices.
It could also suppress innovation and broader economic benefits stemming from AI.
The CMA's Update Paper revealed an "interconnected web" of more than 90 partnerships and strategic investments involving these same companies.
While acknowledging the substantial resources, expertise, and innovation potential these large corporations possess, the CMA cautioned against allowing these powerful partnerships and integrated entities to diminish the competitive capabilities of rival companies or to serve as shields against competition.
The Update Paper highlights three key risks to fair and open competition stemming from the dominance of GAMMA:
- Companies controlling "critical inputs" essential for developing FMs, potentially enabling them to limit access and establish a competitive advantage.
- Large tech companies leveraging their dominant positions in consumer or business markets to manipulate choices for GenAI services and hinder competition in deploying these tools.
- Collaborations among major players that could amplify existing market power positions across the entire value chain.
The paper outlines how the CMA's principles will address each risk and details the actions the CMA is currently taking and considering in the near future to tackle these issues.
The update comes nearly a year after the CMA started an initial assessment of the high-end AI market last May.
Subsequently, the CMA released a set of principles for "responsible" generative AI development, which it stated would shape its supervision of this rapidly evolving market.
In December, the CMA said it was scrutinising the alliance between Microsoft and OpenAI to assess the potential implications of a merger that could transform the AI landscape. The regulator said it was inviting feedback from both companies as well as other interested parties.
Microsoft has made substantial investments in OpenAI, holding a 49% stake in its for-profit subsidiary. The collaboration has evolved over time, with a $1 billion investment in 2019 and an additional $10 billion this year, solidifying Azure's position as OpenAI's exclusive cloud provider.
However, the specifics of their partnership remain largely undisclosed.