IBM’s chief executive told Bloomberg that soon some 7,800 roles might be replaced by AI. The following year Duolingo cut a tenth of its contractor workforce to make room for AI, while Atlassian followed suit. Klarna said its AI assistant does work of 700 customer‑service staff, guiding a headcount goal under 2,000. Jack Dorsey has pledged to keep Block headcount flat while AI drives growth.

A spring 2024 experiment in the UK asked roughly 300 writers to craft short fiction, half with GPT-4 assistance. Independent judges rated the AI‑aided stories as more creative on average, confirming the expected individual boost. However, when researchers examined the full collection, the AI‑generated tales clustered tightly, showing far less stylistic diversity than the unaided pieces.

Separate work published in *Nature* demonstrated that language models trained on recursively generated text begin to “collapse,” losing rare viewpoints and edge‑case knowledge as each generation narrows the distribution. A follow‑up analysis warned that reliance on AI‑summaries skews public belief, pushing collective understanding further from ground truth. The erosion of minority expertise mirrors the convergence observed in the writing study.

The findings suggest a paradox: AI can amplify individual output yet compress collective intelligence. IBM has already retreated from earlier redundancy ambitions, hinting that firms which preserve, rather than replace, human interaction will retain the diversity essential for robust AI systems.