A Visual Studio Code pull request became the focus of a dispute over how an editor should disclose AI involvement in source-code commits. The change set the Git extension's `git.addAICoAuthor` configuration default to `all`, enabling a `Co-authored-by` trailer when the software detected AI-generated code contributions.
Commit trailers are durable metadata. They travel with a Git commit and can be displayed by hosting services, so an automatic attribution is more consequential than a temporary badge inside an editor. The proposed behavior was intended to identify AI assistance, but comments on the pull request argued that it could attach Copilot's name when users did not believe the tool had participated in the work.
Several commenters reported seeing the trailer despite having disabled AI features or not intentionally using Copilot for the affected changes. Those reports are user accounts in the discussion, not a comprehensive technical test across Visual Studio Code versions and configurations. They nevertheless exposed a gap between the feature's stated condition—detected AI-generated contributions—and the behavior some developers experienced.
The pull-request discussion also identified a configuration mismatch. The contributed settings schema had been changed to make `all` the default, while a runtime call in the Git repository code still supplied `off` as its fallback. A reviewer warned that the two values could diverge in environments where the contributed configuration defaults were unavailable, including some tests or hosts, and suggested either matching the runtime fallback or relying on the schema.
Critics raised two distinct objections. The first concerned consent and visibility: users said an editor should not silently modify a commit message after they had written it. The second concerned meaning. If an AI co-author trailer appears too broadly, it can become noisy metadata and may fail to distinguish genuine model contributions from ordinary use of an editor that happens to include AI capabilities.
A maintainer later thanked participants for feedback, apologized for what was described as a regression and said a fix would be pursued for version 1.119. The discussion also pointed readers to a separate update for further detail. The supplied evidence does not establish the final merged behavior or the contents of that later analysis.
The episode illustrates a design challenge for development tools. AI attribution can improve transparency only if the trigger is accurate, the user can see what will be written and disabling the related feature reliably disables the metadata in every commit path. Otherwise, a label meant to clarify authorship can create a new dispute about it.
