Highlight

In a few years, AI experiments running on “virtual cells” may produce results that no human being can explain. The AI may tell us, “Altering gene X while adding drug Y will produce the following physiological changes,” but no one ([not even the AI](https://www.quantamagazine.org/is-ai-reasoning-right-for-the-wrong-reasons-20260731/)) will understand why.[1](https://charleseisenstein.substack.com/p/the-lost-subject-ai-and-the-human?selection=449a47b8-f641-47d3-b5ad-d8e50518d27a&utm_campaign=post-share-selection&utm_medium=web&aspectRatio=instagram&textColor=#ffffff&bgImage=true#footnote-1) No understanding will be necessary to produce useful results. The young college student wonders whether she will be able to find a job.