Generative AI’s systemic extraction: How corporate tech commodifies art, labor, and culture under extractive capitalism

Indigenous epistemologies treat art as living cultural heritage, governed by relational ethics that prohibit its extraction for profit. AI’s training on Indigenous designs, stories, and symbols without consent replicates settler-colonial patterns of cultural appropriation, reducing sacred knowledge to training data. Protocols like the *Māori Intellectual Property Principles* or *UN Declaration on the Rights of Indigenous Peoples* (UNDRIP) explicitly reject such commodification, yet are ignored in AI development.

The ‘art heist’ framing echoes centuries of cultural plunder, from the Napoleonic looting of European art to the 19th-century ‘curiosity’ trade in Indigenous artifacts. Photographic reproduction in the 19th century similarly disrupted art markets, but lacked AI’s scalability and opacity. The enclosure of art via copyright law—accelerated by the Berne Convention (1886)—laid the groundwork for today’s AI-driven appropriation, where corporations claim ownership of derivative works.

In Japan, the concept of *mottainai* (regret over waste) critiques AI’s resource-intensive training, while *manga* artists resist AI tools that undercut their livelihoods. African artists like Wangechi Mutu and El Anatsui have long navigated colonial-era appropriation, now facing AI-generated ‘homages’ that dilute their cultural authority. In Latin America, the *buen vivir* movement rejects AI’s extractive logic, advocating for communal cultural sovereignty over digital tools.

Generative AI’s training data is often scraped without consent, violating copyright law and ethical norms in data collection. Studies show that AI-generated art lacks originality, relying on statistical patterns in training data rather than creative synthesis. The energy and water demands of data centers—e.g., Microsoft’s 2023 report on AI’s water usage—further compound the environmental cost of this ‘heist.’

Artists describe AI as a ‘vampire’ that drains the lifeblood of human creativity, reducing art to algorithmic mimicry devoid of soul. The spiritual dimension of art—its role in healing, ritual, and collective memory—is erased when works are reduced to training data. Creatives like Refik Anadol and Ian Cheng critique AI’s lack of intentionality, framing it as a hollow simulacrum of human expression.

Without intervention, AI could collapse art markets by flooding them with derivative works, devaluing human labor and eroding cultural diversity. Scenario modeling suggests that by 2030, 60% of commercial art could be AI-generated, with artists pushed into precarious ‘prompt engineering’ roles. Regulatory sandboxes in the EU and Brazil offer pathways to rebalance power, but require global coordination to prevent jurisdictional arbitrage.

Global South artists and artists of color are disproportionately affected, as their works are scraped without compensation while Western artists’ styles are monetized. Disabled artists, who often rely on analog tools, face exclusion from digital art markets dominated by AI. Women artists—already underrepresented in galleries—see their styles co-opted by AI tools trained on male-dominated datasets, reinforcing patriarchal norms.