The cyborg is always political. This is one of Haraway's most insistent claims, and it is the claim that popular readings of the Manifesto most consistently miss. The cyborg is not a celebration of human-machine fusion in the abstract. It is a figure for analyzing the specific power relations that constitute every hybrid — who controls the hybridization, who benefits from it, who bears its costs, whose interests the hybrid serves, and whose interests it marginalizes.
The Orange Pill contains a political analysis, but it is primarily economic — focused on who gets access to the tools, who captures the productivity gains, who bears the transition costs. Haraway's politics goes deeper than distribution. It asks about constitution — about the forces that shape the hybrid itself, that determine what the cyborg can do and what it values, whose knowledge it carries and whose it erases.
The most consequential political question about AI is not who gets to use the tools. It is who built them, on whose data, encoding whose values, optimizing for whose workflows, and serving whose interests. The AI systems that constitute the cyborg condition are products of a specific industry concentrated in a handful of companies in a handful of countries. The training data reflects the knowledge architecture of the English-speaking internet. The design choices reflect the priorities of Silicon Valley. The alignment values are chosen by specific researchers at specific companies.
The 2025 AI & Society paper proposing the cybork — cyborg plus work — extended this analysis by arguing that intelligence in AI lies not in any function of isolated systems, but rather in the situated context of their use. The question shifts from "Where is the intelligence of AI?" to "Where does AI intelligently operate?" — shifting attention from the machine's internal properties to the social and political context within which the machine is deployed.
The democratization of capability argument — the claim that AI tools lower the floor of who gets to build — must be examined through the lens of power, not merely access. The student in Dhaka accessing Claude Code is accessing a system whose knowledge architecture was not designed with her in mind. She can build with the tool. But the tool shapes what she can build, what questions she can ask, what solutions she can imagine. The democratization of access is real. The democratization of the terms on which the hybridization occurs is not.
The political analysis has been central to Haraway's work since the 1985 Manifesto, which was published in Socialist Review and was explicitly pitched as socialist-feminist intervention. The extension of the analysis to contemporary AI systems has been developed by scholars including Ruha Benjamin, Safiya Noble, Timnit Gebru, Kate Crawford, and the growing field of critical AI studies.
Access is not governance. Being able to use a system is not the same as having a voice in how the system is constituted.
Constitution is political. Design choices, training data, and alignment values encode contested political decisions presented as technical necessities.
Concentration is the default. Without deliberate intervention, the benefits of hybridization concentrate among those who build, and the costs fall on those who use.
Accountability requires participation. The people affected by a system must have a voice in its governance, not merely access to its outputs.
The priesthood model is insufficient. Better-informed builders making decisions on behalf of a bewildered public is not the answer; democratic participation in governance is.