Elon Musk’s xAI leverages OpenAI’s models via distillation, exposing AI industry’s extractive knowledge monopolies and regulatory gaps
Original framing: “Elon Musk confirms xAI used OpenAI’s models to train Grok” — The Verge
The practice of model distillation echoes historical patterns of corporate appropriation of open-source and academic research, such as IBM’s early AI projects or Google’s acquisition of DeepMind. Silicon Valley’s 'open' innovation rhetoric has long been a Trojan horse for monopolistic control, as seen in the Linux kernel’s co-optation by tech giants. This case is a microcosm of a century-long trend where proprietary systems extract value from public or communal knowledge.
The case of xAI’s use of OpenAI’s models via distillation is not merely a technical dispute but a symptom of a deeper crisis in AI governance, where knowledge is treated as a private commodity rather than a public good.
This crisis is rooted in Silicon Valley’s extractive ethos, which has historical precedents in the co-optation of open-source and academic research by corporate interests. The lack of consent from marginalised communities and the absence of regulatory oversight further entrench these power asymmetries, risking a future where AI knowledge is controlled by a handful of tech giants. Indigenous and Global South epistemologies offer critical alternatives, emphasizing relational knowledge and communal ownership, while trickster figures like Hermes expose the absurdity of corporate 'open' rhetoric. The path forward requires breaking up AI monopolies, mandating open data commons, and co-designing AI systems with marginalised communities to ensure that knowledge remains a tool for liberation, not exploitation.