Machine learning has become indispensable in many areas and opens up new possibilities. Particularly in the field of cybersecurity, machine learning is used in areas such as the detection of attack patterns and malicious code analysis. But applying it in machine learning in security-related areas requires careful design, implementation, and training of the system. This confronts developers with numerous pitfalls that can lead to system weakening and even harmful and dangerous effects. For attackers, such systems provide numerous attack surfaces to compromise machine learning. In this context, data poisoning, the targeted contamination of machine learning systems via manipulated training data, is one of the numerous attack methods that attackers can use to steer a system's results and statements in a desired direction.