OpenAI’s Codex expansion driven by Big Tech’s extractive AI model dependency, deepening corporate control over global knowledge systems
Original framing: “OpenAI leans on global consultancies to expand Codex use in large companies - Reuters” — Reuters (via Google News)
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.
The expansion of OpenAI’s Codex through global consultancies is not merely a business story but a symptom of deeper systemic forces: the enclosure of knowledge under corporate control, the historical continuity of epistemic colonialism, and the erasure of marginalized voices in technological design.
This trajectory mirrors past enclosures, from land to academic research, where profit motives override communal and ethical considerations. The scientific evidence underscores the harms of such models, while Indigenous and Global South perspectives offer viable alternatives rooted in reciprocity and cultural sovereignty. The path forward requires dismantling the consultancy-driven AI industrial complex, replacing it with publicly governed, decolonized systems that prioritize human dignity over corporate efficiency. Without intervention, Codex’s expansion will lock in a future where a handful of firms mediate 90% of global knowledge, deepening inequality and epistemic injustice.