Blog Post

Educator Developer Blog
3 MIN READ

GitHub Copilot Update: New AI Model That Also Filters Out Security Vulnerabilities

AnthonyBartolo's avatar
Feb 16, 2023

Artificial intelligence (AI) is put to good use in GitHub Copilot, a powerful Visual Studio Code extension that offers real-time code suggestions to coders. GitHub Copilot has been evolving and growing in quality ever since its release, and it now features an enhanced AI model, a new vulnerability filtering system to detect insecure coding patterns and additional functionalities. In this post, we'll have a look at the newest updates to GitHub Copilot and discuss how they can help students learn to code.

 

Better AI-powered code suggestions

The latest update's major goal is to increase the quality and responsiveness of code suggestions. The underlying Codex model has been improved to accomplish this. The revised model provides more accurate and faster code suggestions to developers, resulting in GitHub Copilot generating an average of 46% of a developer's code across all programming languages. 
 
GitHub Copilot providing code suggestions

 

The improvements to GitHub Copilot’s AI model include an upgraded Codex model, better context understanding, and a lightweight client-side model. The improved Codex model produces better code synthesis outcomes. A new paradigm known as Fill-in-the-Middle (FIM) allows for better context awareness. FIM provides developers with improved code suggestion prompts by harnessing common code suffixes and allowing a space in the center for GitHub Copilot to fill. As a result, it now has a better understanding of the intended code and how it should fit in with the rest of the program.

 

Additionally, the lightweight client-side model has been changed to increase overall code proposal acceptance rates. It reduces the frequency of undesired suggestions when they may disrupt a developer's workflow by using basic information about the user's context, such as whether the previous proposal was accepted. This resulted in a 4.5% reduction in unwanted suggestions, allowing GitHub Copilot to respond to each developer who uses it more effectively.
 

Harnessing AI to filtering out security vulnerabilities

GitHub Copilot now incorporates an AI-based vulnerability filtering system that prevents insecure code patterns in real-time to make GitHub Copilot suggestions more secure, in addition to the improvements to the AI model. Hardcoded credentials, SQL injections, and path injections are among the most common susceptible coding patterns targeted by the model. With GitHub Copilot, this solution is a huge step in assisting developers in writing more secure code.

 

GitHub Copilot's AI-based vulnerability filtering system is fast and efficient, as it can detect vulnerable patterns in incomplete fragments of code. This is a significant improvement over traditional security vulnerability identification methods, which are often used on entire repositories during build or release with static code analysis tools. GitHub Copilot allows developers to obtain fast, accurate vulnerability detection directly from the editor.

 

GitHub Student Developer Pack

The GitHub Students Developer Pack is a valuable resource for students who are interested in software development. As we've detailed previously, the developer pack gives student free access to hundreds of premium products like GitHub Copilot including all the recent updates. Students can use GitHub Copilot to write code faster, obtain better code suggestions, and learn to create more secure code. The AI-based vulnerability checking system in GitHub Copilot can help students avoid common coding mistakes and build more secure applications. The AI model upgrades and improved code suggestions make it easier for students to write high-quality code.

 

New features in GitHub's Copilot upgrade make the tool even more useful for programmers. While the AI-based vulnerability filtering system makes it simpler to write secure code, the upgraded AI model and better code suggestions will help developers write code faster while filtering out security vulnerabilities. These updates have made GitHub Copilot an even more effective and responsive tool for developers and students to learn new coding languages.

 

Raccoon Engineers updating GitHub Copilot

Updated Feb 15, 2023
Version 1.0
No CommentsBe the first to comment