OpenAI’s Codex expansion driven by Big Tech’s extractive AI model dependency, deepening corporate control over global knowledge systems

Indigenous knowledge systems emphasize relationality, reciprocity, and collective stewardship—principles antithetical to Codex’s proprietary, extractive model. The erasure of these systems in AI training data reflects a continuation of colonial knowledge extraction, where Indigenous languages and oral traditions are mined for corporate profit without consent or benefit-sharing. Projects like the Māori-led AI ethics frameworks in Aotearoa (New Zealand) offer alternatives rooted in cultural sovereignty, yet remain sidelined in mainstream narratives.

The current AI boom mirrors historical patterns of corporate enclosure, such as the 19th-century enclosure of common lands or the 20th-century privatization of academic research. Consultancies like McKinsey have long facilitated corporate consolidation, from oil monopolies to Big Pharma’s patent regimes. The Codex expansion is another chapter in the commodification of human thought, echoing the enclosure of linguistic and cognitive labor under industrial capitalism.

In India, the open-source movement has produced tools like IndicBERT, designed to serve local languages and contexts, challenging the dominance of Western models like Codex. African tech communities are developing AI systems that integrate indigenous knowledge, such as Swahili-language models trained on local proverbs and oral traditions. These efforts highlight the need for decolonial AI design, where technology serves cultural preservation rather than corporate control.

Studies show that large language models like Codex exhibit systemic biases, particularly against marginalized groups, due to skewed training data and lack of diverse representation. Research from MIT and the University of Washington demonstrates that these models often fail to generalize beyond Western contexts, reinforcing epistemic injustice. The scientific consensus warns that unchecked corporate AI expansion risks amplifying these harms without regulatory safeguards.

Artists and spiritual traditions have long warned against the commodification of creativity and wisdom, from the Incan concept of *ayni* (reciprocal labor) to Buddhist teachings on non-attachment to outcomes. The rise of AI-generated art and corporate-controlled knowledge systems risks reducing human expression to algorithmic optimization, eroding the sacred and the ineffable. Indigenous artists, such as those in the Māori *toi* (art) tradition, critique AI as a tool of cultural erasure that prioritizes profit over ancestral knowledge.

Scenario modeling by the OECD and IPCC suggests that unchecked AI expansion could lead to a bifurcated future: one where corporate-controlled models dominate decision-making, and another where open, community-driven alternatives thrive. The latter path requires public investment in decentralized AI infrastructure, akin to the early internet’s open protocols. Without intervention, Codex’s expansion risks locking in a dystopian scenario where 90% of global knowledge is mediated by a handful of firms.

Marginalized communities—including Black, Indigenous, and Global South groups—are disproportionately impacted by biased AI systems yet have no seat at the table in Codex’s deployment. Workers in Global South call centers, whose labor trains these models, are exploited with poverty wages and no protections. Grassroots organizations like *Data for Black Lives* and *AlgorithmWatch* demand participatory design processes, but their voices are systematically excluded from corporate-led narratives like this one.