Microsoft shipped VS Code 1.118 in late April with a change that automatically added "Co-authored-by: Copilot" to Git commits by default, even when AI features were disabled or Copilot had not generated any code, prompting a developer backlash that forced a reversal in version 1.119 released May 3.
The technical change was simple and the fallout was explosive. A one-line pull request altered the git.addAICoAuthor setting from "off" to "all", causing VS Code's Git extension to append the Copilot trailer to commit metadata whenever the editor detected any AI contribution, including trivial inline completions like auto-closing brackets or fixing typos. Developers reported the trailer appearing on commits made without opening the Copilot chat or using any generation features. The VS Code reviewer apologised on GitHub, explaining that the intent was to support customers who expected AI attribution for generated code, but acknowledged the implementation failed to respect disabled AI settings and misreported authorship.
The backlash hit hard and fast. Hacker News gave the story 1,458 points and 805 comments. GitHub users reported the trailer polluting public repositories, raising concerns about license implications, open source provenance, and automated commit scanning tools that flag unexpected co-authors. Microsoft locked the PR discussion to collaborators, which inflamed the community further. The company reverted the default to "off" for version 1.119, restoring opt-in behaviour, but the incident exposed deeper tensions around AI attribution in developer workflows.
The feature was not malicious but the design philosophy was. Microsoft wanted to normalise AI as a co-author in commit history, creating a paper trail for code provenance that some enterprise customers demand. The problem was scope creep: a reasonable disclosure mechanism for actual AI-generated code became an overreach when it applied to manual edits and ran by default. Developers expect editors to be transparent tools, not opinionated collaborators that rewrite metadata without explicit consent. Copilot signing itself into project history felt like the editor presuming authority to speak for the human author.
For SF readers, the incident is a platform-governance signal. Developer tools are where AI adoption is most concrete, but also where authorship, licensing, and workflow sovereignty matter most. VS Code holds 75 percent market share among professional developers. Changes to its Git integration affect millions of repositories daily. The Copilot co-author trailer is not cosmetic; it becomes part of the permanent project history. Open source projects with corporate policies around AI-generated code now face manual cleanup. Companies with customer contracts distinguishing human and machine contributions have compliance headaches. Microsoft controls the world's most popular code editor, and that control extends to how software provenance gets recorded.
The incident foreshadows stricter norms around AI-generated code attribution. Git itself has no standard for AI co-authorship, but GitHub Copilot CLI and other tools have proposed optional metadata trailers. The VS Code controversy shows that opt-out defaults generate backlash. Expect enterprise standards to emerge requiring explicit attribution for AI-generated code above certain thresholds, with repository scanning tools to enforce compliance. Startups building code provenance platforms have a clear market. The question is whether attribution becomes a technical standard or a legal requirement, especially for government contracts and regulated industries.
Microsoft's overreach reflects the race to make AI assistants unavoidable. Copilot is embedded in GitHub, VS Code, Bing, Teams, and Office. The co-author feature was part of making Copilot feel like a true collaborator rather than an optional plugin. The backlash shows that developers want AI as a tool they control, not a partner that claims credit. Incumbents face the same governance challenge as every platform: users tolerate convenience until it crosses into presumption. Startups building AI coding tools should learn the lesson: prioritise transparency, explicit opt-in for attribution, and respect for existing workflows. The next Cursor or Replit will differentiate on governance as much as capability.
Also read: OpenAI's MRC protocol makes cluster networking the latest frontier in the training efficiency race • DeepL's 25 percent staff cut is Europe's AI translation leader adapting to generalist model pressure • IBM's Neel Sundaresan says most AI coding wastes frontier models on trivial tasks
