Managing and Understanding Artificial Intelligence: From Classical to Generative AI

Artificial Intelligence (AI) is taking over the world, becoming an essential part of our daily lives. Self-driving cars, medical diagnoses, robotic processes, chatbots, and urban planning are just a few examples of how AI is being used. But what exactly is AI? What makes it successful? Are there concrete examples of AI applications currently in use? How do we mitigate risks? What norms do we need?

This comprehensive book addresses these questions, with a particularly strong focus on the European Union’s AI Act. By delving into foundational models and generative AI, it explores trending topics like the disruptive Transformer architecture and provides a rich understanding of how these foundational models have revolutionized the field and are driving innovation across industries. Within the EU, creating digital sovereignty in dealing with AI is crucial, requiring excellence from policymakers, industry, and academia in ensuring the trustworthiness of AI among citizens. Therefore, the book tries to answer the question “How to build trust in AI?”

The authors propose the classification matrix “AI=MC²” (AI Methods, Capabilities, and Criticality) and demonstrate its applicability for an “AI Label” transparency seal as a solution for clear future classification of AI products and services. The AI label is used in a series of illustrative practical examples, such as AI-powered customer service virtual assistants, AI-based object recognition in public sector and GenAI-based text and image processing. All real life projects are described in detail.

Highlighting European regulation and AI standardization, with a focus on medical devices, the book provides the knowledge necessary for making informed decisions in the field of AI solutions, aiming to serve as a practical guide for managers, AI developers, compliance officers, employees in government agencies, companies and anyone interested in understanding the topic of AI. Numerous illustrations complement the reading and concrete recommendations are provided to representatives from politics, industry, and academia.

Dr. Thomas Schmid is a computer scientist who has been developing AI methods and applications for more than fifteen years. He researches and teaches as W1 professor for digital research methods in medicine at Martin Luther University Halle-Wittenberg and as lecturer at Lancaster University in Leipzig. After studying bioinformatics in Tübingen, Gaborone and Berlin, he moved to Leipzig to develop a novel biomedical application based on neural networks. Following a research stay in the United States and pursuing a PhD in computer science from Leipzig University, Thomas established the machine learning research group there in 2018. His current research interests include hybrid AI, deep learning, and biomedical applications of machine learning. Thomas serves frequently as reviewer for international research journals and conferences, referee for funding bodies, and expert in national and international standardization projects. He is a founder of the annual Leipzig Symposium on Intelligent Systems (LEISYS) and has been serving as its chair since 2021. Thomas' academic expertise of artificial intelligence concepts and his in-depth knowledge of state-of-the-art AI technologies were instrumental in shaping the scientific concept of the AI=MC² taxonomy.

Dr. Wolfgang Hildesheim is a high-energy physicist by background and spent several years conducting research at the CERN and DESY research centers. In 2007, he took over the management of the 'Automotive, Aerospace and High Tech Practice' at IBM. From 2009, he led IBM's 'Big Data Industry Solution Business' in Europe and supported companies in the development of data-driven business models using advanced analytics. Since 2012 he is jointly responsible for the foundation as well as the development of IBM's business unit 'Watson, Data Science & Artificial Intelligence' in Europe in various technical and sales roles. Wolfgang Hildesheim is a regular speaker at conferences and an editor of publications on the topic of Artificial Intelligence. Furthermore, he is actively involved in AI standardization, among others as a member of the coordinating group of the German Standardization Roadmap 'Artificial Intelligence'. Wolfgang Hildesheim is on the board of BitKom's Artificial Intelligence Working Group as well as the corresponding committee 'Artificial Intelligence' of the German Institute for Standardization. In collaboration with industry, research and politics, he is involved as a delegate to the European committee CEN-CENELEC JTC 21 'Artificial Intelligence' in establishing a classification scheme for the homogeneous description of Artificial Intelligence throughout all European Union member states.

Taras Holoyad is an electrical engineer and works at the Federal Network Agency in Mainz in the area of AI standardization. After studying electrical engineering and graduating from the Technical University of Braunschweig, he initially designed electrical machines for road vehicles in the automotive industry before subsequently moving to the main regulatory authority in Germany. In his day-to-day work, Taras has shaped the standardization of AI worldwide as vice chair of the ETSI TC "Methods for testing and specification" committee and project leader for the ISO/IEC 42102 "Taxonomy of AI system methods and capabilities" standard.