Large language models may homogenize human expression; resistance varies by cognitive diversity
Original framing: “AI can 'same-ify' human expression — can some brains resist its pull?” — Nature
The homogenization of language through AI echoes historical patterns of linguistic imperialism, such as the suppression of indigenous languages during colonial rule. These parallels show how technology can be a continuation of older power structures.
The systemic challenge of AI-driven homogenization of human expression is deeply intertwined with historical patterns of linguistic suppression and contemporary power structures in the tech industry.
Indigenous and non-Western linguistic practices offer alternative models of resistance and resilience, while scientific analysis reveals the biases embedded in AI training data. Marginalized voices are systematically excluded from AI development, reinforcing the very homogenization the technology risks enabling. To counter this, a multi-pronged approach is needed: diversifying training data, promoting ethical governance, and supporting community-led AI initiatives. These steps can help preserve cognitive and linguistic diversity in the face of algorithmic standardization.