Hardly any other technology has produced as many groundbreaking advances in recent months – we are talking about artificial intelligence (AI).

One example from the field of medical software development is GitHub Copilot. As an AI tool, it supports software developers in programming. Copilot learns from the public code repositories on GitHub and can, for example, make code suggestions, complete them and generate entire boilerplates.

BAYOOMED starts tests with GitHub Copilot for medical software development

We at BAYOOMED wanted to find out precisely: Do AI tools like GitHub Copilot deliver what they promise? Can we implement software projects more efficiently with Copilot and what are the weaknesses of the AI tool?

For our trial, two teams worked on different projects during a one-month test phase. Team A, which focused on native mobile development with Android/Kotlin and iOS/Swift with new UI frameworks, found that Copilot’s suggestions were often unhelpful and not aligned with the use of the frameworks. The tooling for Copilot in XCode was also less advanced.

In contrast, Team B’s assessment was positive. They were working on a cross-platform application with C#/.NET Framework. After a short learning phase, GitHub Copilot’s suggestions accelerated their development, especially for repetitive tasks such as writing tests.

Conclusion

Is GitHub Copilot really helpful for software developers when programming?

In summary, we can confirm that GitHub Copilot can significantly improve the software development workflow, but its effectiveness depends on the amount of work involved. Copilot works best when the AI tool is used for programming already familiar tasks. The weakness of GitHub Copilot, on the other hand, clearly lies in working with unknown frameworks.

And without our experts at BAYOOMED, it wouldn’t work at all. GitHub Copilot does not require tool validation, as our software developers review and validate Copilot’s code proposals and output to ensure accuracy, quality and security before deploying the code into their project. GitHub Copilot is not able to independently generate essential parts of the software without developers compiling and integrating the code.

Bernd Seidenspinner
CTO at BAYOOMED

Our test run with Copilot was accompanied by BAYOOMED CTO Bernd Seidenspinner. Bernd has been part of BAYOOMED for more than 15 years and is our team’s permanent contact for all questions relating to new technologies in medical software development.