US AI giants unite to monopolize global model access amid fears of Chinese innovation diffusion and open-source erosion

Indigenous knowledge systems universally treat innovation as a communal process, where tools and ideas are shared to benefit the collective rather than hoarded for profit. This contrasts sharply with the proprietary model of AI development, where models are locked behind paywalls or corporate firewalls, denying access to researchers in the Global South or marginalized communities. The Western framing of 'cloning' as theft ignores how indigenous societies have long engaged in iterative, collaborative knowledge-building without legal restrictions.

The current AI monopoly crisis mirrors historical patterns of corporate consolidation, such as the railroad trusts of the 19th century or the oil monopolies of the early 20th century, where a few firms controlled critical infrastructure to extract rents. The open-source movement in software (e.g., Linux, Apache) emerged as a direct response to such monopolies, yet today’s AI giants are repeating the same cycle by co-opting open-source tools while restricting access to foundational models. The 1980s US antitrust cases against IBM and AT&T offer a precedent for breaking up monopolies, but today’s regulators appear paralyzed by the complexity of digital markets.

Cross-culturally, the tension between proprietary and communal knowledge is stark: while Silicon Valley treats AI models as trade secrets, communities in India, Africa, and Latin America have developed open, decentralized AI tools (e.g., Indic language models, Swahili NLP projects) to serve local needs. The Chinese government’s 'open innovation' policies, despite state capitalism, have enabled rapid AI adoption in sectors like healthcare, contrasting with the West’s reliance on closed systems. Even in Japan, companies like Sony have historically balanced proprietary innovation with open collaborations, offering a middle path.

Scientifically, the claim that 'cloned' models pose a threat is dubious: AI models are statistical approximations of data, and replication is a core part of scientific validation. Studies show that open-source models often outperform proprietary ones in niche applications due to community-driven fine-tuning (e.g., StarCoder, Bloom). The real issue is not 'cloning' but the concentration of computational power: the top 10 AI labs control 90% of GPU resources, creating an artificial scarcity that stifles innovation. This mirrors the replication crisis in psychology, where closed research practices led to flawed findings.

Artistically and spiritually, the proprietary AI model reflects a mechanistic worldview that divorces knowledge from its social and ecological context. Indigenous and Eastern spiritual traditions (e.g., Daoism, Shinto) view technology as an extension of nature, requiring harmony with community and environment. The Western obsession with 'owning' AI models ignores how creativity thrives in collaborative, non-hierarchical spaces, as seen in open-source art projects like Stable Diffusion’s community-driven iterations. This framing also overlooks the spiritual dimensions of AI, where models are seen as 'living' entities evolving through collective interaction.

Future scenarios suggest that the current alliance will accelerate a bifurcated AI ecosystem: a proprietary 'premium' tier dominated by US firms and a fragmented 'open' tier with limited resources, widening the global digital divide. If unchecked, this could lead to a 'feudal AI' system where corporations act as gatekeepers to critical infrastructure, with users as serfs paying rent for access. Conversely, a shift toward open, federated models could democratize AI, enabling Global South researchers to develop culturally relevant tools. The EU’s AI Act and India’s open-source push offer potential counter-models to this dystopian trajectory.

Marginalized voices—particularly researchers from Africa, Latin America, and Indigenous communities—are systematically excluded from the AI development pipeline due to lack of access to proprietary models and computational resources. The narrative of 'Chinese cloning' further marginalizes these groups by framing them as passive recipients of Western innovation rather than active contributors. Initiatives like Masakhane (African NLP) or the Indigenous AI Ethics Council highlight how these communities are building alternatives, but their work is sidelined in mainstream discussions. The US firms’ alliance risks entrenching these inequities by reinforcing a top-down, corporate-controlled AI ecosystem.