Nowadays, various applications of artificial intelligence (AI) are clearly seen in many fields, therefore, it may seem like a technological achievement of the 21st century. However, its origin dates back to the middle of the last century.
The questions arise as to how AI can be illuminated from the perspective of data protection, and especially what the main data protection concerns are, moreover, what kind of data protection risks it poses, and what practical solutions are known about the topic. The purpose of this paper is to provide an insight into the data protection approach to AI through some practical examples.
Based on the results it can be established that this futureproofing technological solution poses several challenges for data protection. As a learning algorithm based on poor foundations can also lead to erroneous conclusions, AI requires a fair amount of appropriate data in order to provide reliable results. It should also be highlighted that profiles can be created from enormous amounts of data and conclusions may be drawn about our habits, which raises concerns. Additionally, its reliability is in question, not only due to the basic data, but due to the self-learning, “black-box” system. The knowledge on which it bases assertions about “something” is very limited. It is obviously a high risk from a data protection perspective.