AI now writes a big share of the code we ship. Should a human have to review every line of it?

Coding assistants now write a large share of new code at many companies - including code that runs banks, hospitals and UPI apps. When that code breaks, nobody can say "the AI did it".

Tools such as GitHub Copilot, Cursor and Claude now suggest whole functions while a developer types, and Indian IT services firms and start-ups have rolled them out to large parts of their engineering teams. A growing share of each day's code starts as an AI suggestion that a developer accepts, edits or rejects.

How closely that code is checked varies. At some companies every change, whoever or whatever wrote it, needs a person to read it and approve it before it ships. Others let automated tests and security scanners clear small or low-risk changes, and keep human review for the rest.

The question here is narrow: should a person have to read and sign off on every line an AI writes before it goes live?

Where do you stand? Vote, then share your opinion - and say what would change your mind.