The impact of AI tools on creating a web page

Abstract

Advances in Artificial Intelligence have drawn huge global attention, namely the recent release of ChatGPT, Bard, and other large language model (LLM)-based chatbots. Today, software developers can use AI to create code, find bugs, set up test suites, and generate test and sample data. The recent development of Artificial Intelligence has enormously impacted the field of programming, fantastically increasing productivity and facilitating more of the hard work of programmers. In general, experts agree that the AI field is the most promising for the future of the computing area from all the recent technologies, and will strongly impact many other areas as well, changing the way of working in several jobs. Companies must study and implement these new technologies in order to remain competitive. This dissertation presents an experiment where two independent developers, from the Portuguese company ARMIS, build a particular web page, one using AI tools such as GitHub Copilot, Builder.io, etc, and the other one doesn't. The objective is to measure the real gains in productivity and code quality. The programming languages and frameworks involved in the project were mainly TypeScript, Angular, Chart.JS, and Axios. This research will analyze the productivity impact of the adoption of these AI tools, as well as code quality and code security. The feedback from the developers and the project managers from ARMIS will also be analyzed. At the end of this dissertation, ARMIS will have a clear view of the potential of AI technologies in its works, and will be able to improve their way of working in many areas.