GitHub disables Copilot ads in pull requests after developer backlash
GitHub has disabled Copilot's ability to insert promotional tips into pull requests following developer backlash. The feature, which injected Raycast ads into over 11,400 pull requests without explicit developer consent, was enabled when Copilot was mentioned in PRs it didn't create. GitHub's product manager acknowledged the decision was a mistake.
GitHub has disabled Copilot's controversial feature that injected promotional tips into pull requests, acknowledging the decision to let the AI tool modify PRs it wasn't created by was "the wrong judgment call."
The issue surfaced Monday when Australian developer Zach Manson discovered that Copilot had inserted a Raycast advertisement into one of his pull requests after a coworker asked the AI to fix a typo. The promotional message read: "Quickly spin up Copilot coding agents from anywhere on your macOS or Windows machine with Raycast," complete with a link and lightning bolt emoji.
Scale of the Problem
Manson's discovery revealed the feature was widespread: over 11,400 pull requests contained identical Raycast tips, with additional PRs containing different promotional messages. The ads appeared to be automatically inserted whenever Copilot was mentioned in a pull request, regardless of whether the AI tool had created the PR.
"I wasn't even aware that the GitHub Copilot Review integration had the ability to edit other users' descriptions and comments," Manson told The Register. "I can't think of a valid use case for that ability."
GitHub's Response
By Monday afternoon, GitHub moved to disable the feature. Tim Rogers, principal product manager for Copilot at GitHub, posted on Hacker News that the feature was intended to "help developers learn new ways to use the agent in their workflow." However, he acknowledged that allowing Copilot to modify pull requests written by humans without their knowledge crossed a line.
"On reflection, [it] was the wrong judgment call," Rogers wrote. "We've now disabled these tips in pull requests created by or touched by Copilot, so you won't see this happen again."
GitHub VP of developer relations Martin Woodward clarified on X that Copilot injecting tips into pull requests it creates has been standard behavior. The newly disabled feature was the ability to modify any PR when mentioned, a capability added more recently.
"When we added the ability to have Copilot work on any PR by mentioning it the behaviour became icky," Woodward stated.
What This Means
This reversal illustrates the tension between monetizing AI tools and maintaining developer trust. GitHub was attempting to use Copilot as a marketing channel by inserting sponsored content directly into developer workflows without explicit consent. The swift backlash and quick rollback suggest companies cannot rely on implicit approval for invasive AI features, even minor ones. The incident also raises questions about AI systems having edit permissions on shared development artifacts—a capability that remains problematic even if the ads are removed.
Related Articles
Microsoft expands Copilot Cowork with AI model critique feature and cross-model comparison
Microsoft is expanding Copilot Cowork availability and introducing a Critique function that enables one AI model to review another's output. The update also includes a new Researcher agent claiming best-in-class deep research performance, outperforming Perplexity by 7 points, and a Model Council feature for direct model comparison.
Microsoft Copilot Researcher adds multi-model features using GPT and Claude
Microsoft has enabled its Copilot Researcher tool to simultaneously leverage OpenAI's GPT and Anthropic's Claude through two new features: Critique, which uses GPT responses refined by Claude, and Model Council, which displays side-by-side outputs with agreement/disagreement analysis. Both features are rolling out in the Microsoft 365 Copilot Frontier early access program.
Suno v5.5 adds voice cloning and custom model training for Pro subscribers
Suno has released v5.5 of its AI music generation model, prioritizing user customization over quality improvements. The update includes Voices (voice cloning via user-uploaded recordings), Custom Models (style training on user music catalogs), and My Taste (preference learning), with voice and custom features limited to Pro and Premier subscribers.
Replit Agent 4 overhauls design, collaboration, and build workflows for product teams
Replit has released Agent 4, a major update that fundamentally restructures how product teams design, collaborate, and build. The release replaces Design Mode with an infinite Design Canvas, moves from fork-and-merge to shared real-time collaboration, and enables concurrent planning and execution.
Comments
Loading...