AI 導致電子廢棄物激增
Retirement
Risk Description
Rapid iteration cycles of AI hardware lead to massive electronic waste generation. GPU, TPU, and other AI-specific hardware lifecycles are increasingly shorter as performance improvements in new generations make old equipment prematurely obsolete. Large quantities of AI hardware containing toxic substances (lead, mercury, cadmium) enter waste streams, threatening the environment and health of communities near processing sites.
Framework Mappings
iso 23894R8
tw principle永續發展與福祉
tw risk type社會系統性影響（Societal Systemic）
Controls English translation pending
- 硬體生命週期延長
- 循環經濟策略
Implementation Steps English translation pending
- 延長 AI 硬體的使用生命週期，透過軟體最佳化而非硬體替換提升效能
- 建立 AI 硬體的回收和再利用計畫
- 在採購評估中納入硬體供應商的環境責任紀錄
- 追蹤並報告組織的 AI 硬體廢棄物產生量