AI's visual culture reveals extractive tech paradigms: how generative imagery obscures labor, identity, and systemic power in Silicon Valley's gaze

Indigenous epistemologies view art as a living, relational practice tied to land, ancestry, and communal responsibility, not a commodity to be scraped and repurposed. The New Yorker's illustration reflects a Western colonial gaze that separates the artist (Altman) from the cultural labor of creation, erasing the Indigenous concept of 'art as stewardship.' Generative AI's training on Indigenous motifs without consent or attribution mirrors historical patterns of cultural appropriation, such as the commodification of Native American designs by Western fashion industries.

The fetishization of the 'lone genius' CEO echoes 19th-century industrial capitalism, where robber barons like Rockefeller and Carnegie were mythologized as self-made visionaries despite exploiting labor and monopolizing resources. The appropriation of artistic styles by tech platforms mirrors colonial-era 'ethnographic' collections, where Indigenous and non-Western art was extracted, categorized, and displayed in Western institutions without context or compensation. The current AI art boom follows a pattern seen in photography and cinema, where new technologies disrupted creative labor before being co-opted by corporate interests.

In Japan, the concept of 'mottainai' (regret over waste) challenges the Silicon Valley ethos of endless extraction, offering a framework for ethical AI development that values resource stewardship. African philosophies like 'Ubuntu' ('I am because we are') reject the individualism embedded in generative AI's 'original' outputs, emphasizing communal creation and interdependence. The Middle Eastern tradition of 'hikma' (wisdom through storytelling) contrasts with AI's reduction of narrative to statistical patterns, highlighting the loss of nuance in algorithmic art.

Generative AI models like DALL-E and MidJourney rely on datasets scraped from the web without consent, often containing biased or mislabeled images that reinforce stereotypes. Studies show that AI-generated art disproportionately favors Western, Eurocentric aesthetics, reflecting the geographic and demographic skew of training data. The energy consumption of training these models (e.g., 1,287 MWh for Stable Diffusion) raises ethical questions about the environmental cost of 'creative' labor replacement.

Artistic traditions across cultures emphasize the sacredness of creation, from the Hindu concept of 'darshan' (seeing the divine in art) to the Christian idea of art as a form of worship—concepts alienated by AI's commodification of imagery. The New Yorker's illustration reduces Altman to a hollow vessel for algorithmic faces, mirroring the spiritual emptiness of techno-utopianism, where meaning is outsourced to machines. Indigenous and folk art often incorporate rituals of permission and reciprocity, absent in the extractive logic of generative AI.

If unchecked, generative AI could lead to a cultural monoculture where corporate-owned models dictate artistic trends, marginalizing diverse voices and reinforcing power imbalances in the creative economy. Scenario planning suggests that without regulation, AI art could exacerbate the precarity of human artists, particularly in Global South markets where labor is already undervalued. Conversely, a future where AI is used to democratize art—with strict ethical guardrails and profit-sharing models—could empower marginalized creators and preserve cultural diversity.

Data annotators in Kenya, India, and the Philippines—often women and gig workers—are the invisible labor behind AI art, paid poverty wages to label images for systems that may replace their jobs. Artists of color and non-Western creators face their work being scraped without consent, while their styles are repurposed in corporate AI outputs without credit or compensation. The New Yorker's illustration, like most AI art discourse, centers white male executives, erasing the global, diverse workforce that makes these systems possible.