BlockBeats 消息，3 月 7 日，Anthropic 最新發布的研究顯示，儘管 AI 理論上可覆蓋商業、金融、法律、計算機等領域的絕大部分任務，但實際採用率僅為一小部分——以 Claude 模型為例，其對計算機和數學崗位的理論覆蓋率達 94%，實際使用僅 33%。研究引入「觀察暴露度」指標，對比理論能力與現實使用數據。
Results show that the group most exposed to AI is not blue-collar workers, but highly educated, high-income female white-collar workers: compared to the low-exposure group, this group has 16 percentage points higher proportion of women, 47% higher average income, and nearly four times higher proportion of individuals with a graduate degree. Researchers warn that as AI capabilities improve and adoption deepens, AI could trigger a “great recession for white-collar workers”—a scenario similar to the unemployment rate doubling from 5% to 10% during the 2007–2009 financial crisis.
目前雖未發生，但風險明確。當前影響更多體現在招聘放緩而非裁員：後 ChatGPT 時代，暴露職業的求職率較 2022 年下降 14%，22-25 歲年輕工作者在相關領域就業率下降 16%。部分年輕人選擇繼續深造或暫避勞動力市場。