Microsoft is deploying its own artificial intelligence to manage a growing wave of browser extensions built using AI-assisted development tools. The company has updated its review pipeline for Edge add-ons to include automated checks that can analyze code generated largely by large language models, aiming to maintain quality standards without delaying submission times.
The Vibe-Coding Shift
The term "vibe-coded" describes extensions written with heavy assistance from large language models such as ChatGPT or GitHub Copilot. Developers prompt the AI to produce functional code, often without fully understanding the underlying logic. This approach has accelerated extension creation but also introduced new risks, including security vulnerabilities, unnecessary permissions and inconsistent behavior. Microsoft's Edge team acknowledged the trend, saying it remains committed to the same quality bar that existed before the era of AI-assisted development.
AI on AI: Microsoft's Automated Checks
Microsoft's response involves integrating machine learning models directly into the Edge extension review pipeline. These models scan submissions for code patterns that violate store rules, flagging anomalies that resemble common AI-generated mistakes. The system prioritizes extensions that appear fully automated, subjecting them to deeper scrutiny before approval. By handling the initial triage with AI, Microsoft hopes to reduce the burden on human reviewers while catching low-quality submissions early. The approach mirrors similar strategies used by other platforms, including Chrome's Web Store, which also grapples with AI-produced add-ons.
Why This Matters
The decision to fight AI-generated code with more AI represents a pragmatic but untested strategy for platform moderation. If the automated checks fail to catch subtle security flaws, users could face compromised browsing sessions or data theft. Developers, on the other hand, may find their efficiently written extensions blocked by opaque algorithmic rules, creating friction in the marketplace. The broader implication is that AI-assisted software production is outpacing traditional review systems, forcing companies like Microsoft to choose between speed and safety. The outcome of this experiment will influence how other platforms handle the coming wave of machine-authored applications.
What It Means for Developers
Extension creators who rely on AI coding tools should expect stricter automated gatekeeping when submitting to Edge. Microsoft recommends testing extensions thoroughly before submission and reviewing permissions against store policy manually. The company also encourages developers to understand the code their AI generates, rather than treating the output as final. For the Edge ecosystem, the new review process could mean fewer low-effort extensions reach users, preserving the store's reputation as a curated marketplace.



