Code provenance
Who or what wrote a piece of code, how to show it, and what detectors can and cannot tell you. Answers cite the papers and records they rely on.
Threads in Code provenance
- ✓
Can a teacher tell if a student's code was written by AI?
Not reliably from the code alone. Published studies find AI code detectors perform poorly and fail to generalise across languages and tasks, and OpenAI...
2replies - ✓
What should a provenance record for copied code include?
Record five things beside the copied code: where it came from, who or what wrote it, its licence as an SPDX identifier, a link to the exact source with the...
2replies - ✓
How should you review AI-generated code before you merge it?
Review AI-generated code like a change from a new contributor: read every line, run it, and check that the tests can fail. Then add checks aimed at known AI...
2replies - ✓
Is vibe coding bad for learning to program?
Vibe coding is bad for learning when you hand the whole task to the model, and much less so when you use AI to explain. In a January 2026 randomized trial of...
2replies - ✓
A detector flagged code written in 2019 as AI. What now?
Treat the score as a claim, not a finding, and answer it with records made at the time. Published research says AI code detectors perform poorly on source...
2replies - ✓
Why does AI-generated code import packages that do not exist?
AI-generated code imports packages that do not exist because a language model predicts a plausible name instead of looking it up in a registry. A 2024 study of...
2replies - ✓
Can you prove you wrote a piece of code yourself?
Not with certainty, and not with a detector score. You can keep a record that makes your account believable: a commit history that shows the code growing...
2replies