The world of software development has fundamentally changed because of the explosive growth of open-source repositories in recent years. Open-source repositories have become a valuable tool for software developers and researchers because they are free and usually easy to use. Likewise, it has been a while since learning Artificial Intelligence (AI) and Machine Learning (ML) skills became very popular among students, especially computer science students, as graduates with these skills are in high demand. Particularly, software engineers are increasingly requiring AI skills to drive innovation, solve complex problems, and remain competitive. There are several open-source projects in the AI/ML space that contain code explanations, e.g., comments and/or documentation, making them potential educational tools. However, it is currently unclear how well AI novices can benefit from these resources. To provide preliminary evidence for the effectiveness of our educational idea, we studied how computer science bachelor students perceive self-study with open-source repositories to build more complex AI/ML systems. After a learning period, we surveyed the perception and learning outcomes from the viewpoint of 112 students. By analyzing the responses, we found that 75% of the students stated that they could now build complex AI/ML systems if provided with enough documentation and descriptions and are motivated to work on them. While this indicates that learning or improving AI/ML skills via open-source repositories is promising, more research beyond self-reporting is needed.