Openness in AI and Downstream Governance: Global Value Chain Analysis Reveals Potential for Catch-up Despite Concentrated Power

The accelerating development of artificial intelligence presents both opportunities and challenges for global economic power, with concerns growing about the concentration of control within a few dominant companies. Christopher Foster, alongside colleagues, investigates this dynamic by examining the phenomenon of ‘openness’ in AI, where companies promote access to models and datasets. This research moves beyond simple descriptions of open AI, instead conceptualising it as a specific type of relationship between firms, allowing the team to apply value chain analysis to understand how foundational AI companies interact with downstream adopters. By mapping these connections, the researchers aim to clarify the capitalist forces at play and predict the types of governance structures that will emerge as AI technologies become more widely integrated into various industries, ultimately offering a more nuanced understanding of AI’s potential for both concentration and diffusion of power.

Global Production and Algorithmic Management

This study consolidates a comprehensive overview of existing research on global production networks, algorithmic management, and the evolving landscape of artificial intelligence, establishing a robust base for further investigation into the intersection of technology, work, and economic power. Researchers systematically examined a wide range of publications to create a comprehensive record of scholarly contributions, providing a valuable resource for scholars and practitioners seeking to understand the complex dynamics shaping the future of work and the role of artificial intelligence in global production systems.

AI Value Chains and Foundational Firm Power

This research pioneers a novel application of Global Value Chain (GVC) analysis to understand the economic dynamics of artificial intelligence and the implications of increasing openness in the field. Researchers systematically mapped the complex relationships within AI production, conceptualising AI as a unique type of interfirm relation, and developed a simplified downstream model identifying “foundational AI firms” as central actors coordinating production and wielding significant power. The team meticulously traced the flow of value from these foundational firms through various stages of production, drawing upon a dataset mapping technical supply chains linked to major foundation models from 2023 to reveal the interplay between open and proprietary toolsets.

Strategic Openness and Freemium AI Models

This work investigates the evolving landscape of artificial intelligence, analyzing the strategic implications of “openness” in AI development. Researchers conceptualise openness not merely as a technical characteristic, but as a specific type of relationship between firms, revealing a spectrum of approaches ranging from fully proprietary models to those with limited releases, categorised as “strategic market openness” when aligned with firm goals. Analysis demonstrates that major AI firms often release less advanced models as a “freemium” offering, attracting users to their platforms, while also driving competition and reducing costs.

Openness Mediates AI Technology Use

This work advances understanding of the rapidly evolving artificial intelligence sector by conceptualising openness in AI as a distinct type of relationship between firms, amenable to analysis through value chain frameworks. By applying this approach, researchers have extended existing maps of AI value chains to link foundational AI development with downstream applications, suggesting that the increasing prevalence of open AI resources may facilitate knowledge transfer, reduce costs, and build skills among actors outside dominant companies. This research demonstrates that openness in AI, even when partial, can mediate technology use and potentially enable competition within the heavily capitalised AI startup landscape. While acknowledging concerns about “openwashing”, the study highlights the potential for broader types of openness to foster learning and innovation, noting that fully assessing the long-term economic impacts of AI openness requires ongoing observation of the sector’s evolution. Future research directions include continued monitoring of value chain dynamics and further investigation into the conditions under which openness can genuinely promote competition and broader societal benefits within the AI ecosystem.