The real problem of our times is not developing more AI but ensuring sustainable futures. Imaginaries of AI for sustainability present logics of progress, smartness, and efficient use of natural resources, whilst rarely acknowledging the environmental impact of these systems. In this chapter, we use the method of defamiliarisation, making the familiar unfamiliar, to explore governance approaches to AI infrastructure that stay within planetary boundaries. We argue that centring nature in AI governance debates opposes Europe’s current state-market entanglements. Rather, it positions the state as the key power-holder, which should wield its economic and technological statecraft for the public interest and counter the accumulation of capital in the hands of the few. This chapter contributes to the understanding of the state’s role in AI infrastructure governance and straddles the fields of infrastructure studies and critical data centre studies.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

More Compute for a Burning Planet? A Scarcity Approach to AI Infrastructures

  • Fieke Jansen,
  • Niels ten Oever

摘要

The real problem of our times is not developing more AI but ensuring sustainable futures. Imaginaries of AI for sustainability present logics of progress, smartness, and efficient use of natural resources, whilst rarely acknowledging the environmental impact of these systems. In this chapter, we use the method of defamiliarisation, making the familiar unfamiliar, to explore governance approaches to AI infrastructure that stay within planetary boundaries. We argue that centring nature in AI governance debates opposes Europe’s current state-market entanglements. Rather, it positions the state as the key power-holder, which should wield its economic and technological statecraft for the public interest and counter the accumulation of capital in the hands of the few. This chapter contributes to the understanding of the state’s role in AI infrastructure governance and straddles the fields of infrastructure studies and critical data centre studies.