Firstly, this Report section raised the unique nature of AI compared to other digital markets due to its distinguished complexity, higher marginal costs, common partnerships, broad applicability, and wide range of users. As such, AI advancements introduce new market dynamics that can either promote or hinder competition. Key submission points include:

• Resource dependence:AI development relies heavily on resources such as computing power, data centers, and specialized chips. These resources are often controlled by major tech firms.
• Data control:The dominance of large incumbents in controlling data can create significant barriers to entry.
• Competitive dynamics:AI introduces new dynamics to competition, such as algorithmic pricing (algorithms are the core building blocks for AI), which can enhance or undermine competitive practices.
• Barriers to entry:The concentration of resources can create barriers for new players and influence the competitive landscape, although synthetic data and innovative training techniques may mitigate some of these barriers.

Secondly, the Report section identified that although the AI landscape has many startups, the industry is dominated by large firms. This extends to all three key layers of AI technology: the Deployment Layer, the Development Layer, and the Infrastructure Layer.

Thirdly, the barriers that new and smaller firms face when entering the AI market was outlined. This was approached through the three critical inputs for AI development: data, compute power, and human expertise. Many Discussion Paper respondents noted that there is a difficulty accessing critical inputs, particularly data and compute, which creates industry involvement barriers for new and smaller firms. While respondents did note that SMEs can leverage compute resources from other firms to save on up-front-costs, large firms still hold high control and market power over the input layers of the AI technology stack. Although these outlined difficulties are prominent, some submissions did provide a counterargument that the AI input market is SME friendly due to strong investments in the sector, an increasing availability of cloud and open-source computing, and promoted partnerships between large and small firms. Furthermore, regarding vertical integration and barriers to entry, suggested potential solutions to foster a more inclusive and competitive AI market were to safeguard access to data, develop public policy, and promote fair partnerships.

Finally, the section noted how AI can assist SMEs through “AI democratization”, which is a process that involves making AI more accessible to a wide range of users who can then leverage the technology to compete more effectively in a wide range of industries. This process would thus strengthen and enable SMEs to compete against larger companies and create a more diverse market. Identified avenues to achieve this process include financial access and efficiency, as well as financial and operational efficiency.