Scholarlink

This chapter claims that two different problems are lurking behind epistemic opacity. One is philosophical, related to skepticism and very hard to solve. The other one is practical, has been solved for different technologies in the past, but for ML the jury is still out. Philosophical treatments have understandably focused on coming up with proposals how to solve the philosophical problem of opacity. This chapter traces how the practical problem has been solved for methods that predated ML and suggests that habituation to an agreed upon level of residual opacity plays a central role. We therefore should expect that habituation to opacity in ML is also possible.

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

How I Stopped Worrying and Learned to Love Opacity

  • 章节

  • 出版模式：

    开放获取

  • 发表日期：

    2026年

查看全文
  • Nico Formánek

摘要

This chapter claims that two different problems are lurking behind epistemic opacity. One is philosophical, related to skepticism and very hard to solve. The other one is practical, has been solved for different technologies in the past, but for ML the jury is still out. Philosophical treatments have understandably focused on coming up with proposals how to solve the philosophical problem of opacity. This chapter traces how the practical problem has been solved for methods that predated ML and suggests that habituation to an agreed upon level of residual opacity plays a central role. We therefore should expect that habituation to opacity in ML is also possible.