Data poisoning is a term from the fields of artificial intelligence, cybercrime and cybersecurity, as well as big data and smart data. It describes the targeted manipulation of data that machines and algorithms use for learning. The goal of data poisoning is to falsify the quality or results of these systems, thereby causing unwanted or even harmful decisions.
Imagine a company using artificial intelligence to sort applications. If someone deliberately injects false or misleading data into the system – for example, manipulated CVs – this can lead to unsuitable candidates being selected as particularly suitable later on. The system is then „poisoned“ and no longer makes reliable decisions.
Data poisoning is a major challenge for companies that use technologies like AI or Big Data. Those who protect their data quality and strictly regulate access rights can better protect themselves from such attacks. Awareness within the company is therefore particularly important so that potential vulnerabilities can be quickly identified and closed.
Deutsch 
English (UK)