GitHub Copilot is an AI service which suggests line completions and entire function bodies as a programmer types. GitHub Copilot is powered by the Open-AI Codex AI system. It is trained on public Internet text and millions of lines of code that are publicly available on websites like StackOverflow, GeeksforGeeks, and many more. While Copilot might be a major time saver that many people would consider “magic” ,it’s also been met with criticism by other developers, who worry that the tool could violate individual users’ copyrights.

How Copilot works?

GitHub describes Copilot as the AI equivalent of pair-programming, which is a term coined for two developers who work together at a single computer. Pair programming is when one developer write code for the problems stated and the other observes and make changes(debugging). In practice, though, Copilot is more of a time saver, which integrates the resources that developers might otherwise have to look up elsewhere. As users type into Copilot, it will suggest lines/blocks of code to add by just clicking a button. That way, they don’t have to spend time searching through various documentations or looking up sample code on sites like StackOverflow.

GitHub also wants Copilot to get more efficient over time based on the collected data from users. So whenever users accept or reject Copilot’s suggestions, its powerful machine learning model will use that feedback to improve future suggestions. This would make Copilot only better with time

Criticism

Not long after Copilot’s launch, many developers started ranting about the use of public code to train the tool’s AI. The major concern being commercial use of open source code without proper licensing. The major reason why developers are criticizing it is: Microsoft, the company that owns GitHub, has access to all the repositories. Training a machine learning model that uses all the public repositories and charges a subscription fee for others to use it will benefit Microsoft. So what do people who contribute to open-source get in return? The tool could also leak personal details that the developers may have posted publicly.

Microsoft’s Policy

The developers, programmers and the open-source community cannot complain, nor can they sue Microsoft. This is because there are absolutely no rules or regulations on how Microsoft plans to use open-source repositories. Even if the open source community decide to sue Microsoft, that would just mean a new set of rules would be imposed on how open-source software is used

The open-source community has mixed feelings regarding Microsoft’s Policy. Some people think that GitHub Copilot doesn’t work the way it’s advertised. A large portion of what Copilot outputs is already full of copyright/license violations, even without extensions. Some think, the code is still AI generated and not a copy-paste block from some repository, so the production is still the programmer’s responsibility.

Conclusion

It’s true that Microsoft is using the public repositories for their own good, but there are no laws based on which people can sue them. This is why most are moving their code from GitHub. It’s indeed copyright infringement for sure, but it’s going to be a little longer before Copilot will deliver a genuine productivity boost. Right now, the suggested snippets do look very accurate, but when dug beneath the surface and you will find that it doesn’t always do what you expected. Can we really find ourselves working with an AI pair-programmer in the future? As for now, it does look skeptical. But with Copilot, the future doesn’t look so far off.