Management von Künstlicher Intelligenz in Unternehmen

Abstract

Zusammenfassung: Künstliche Intelligenz bietet Unternehmen neue Möglichkeiten Prozesse, Produkte, Dienstleistungen und Geschäftsmodelle zu innovieren und bestehende zu verändern. Daher wird das professionelle Management Künstlicher Intelligenz in Unternehmen zu einer zentralen Aufgabe, um die neuen Wertversprechen mit produktiven Systemen zu realisieren. Der Beitrag stellt das St.Galler Management Modell für KI (SGMM-KI) vor und zeigt sieben Handlungfelder für den betrieblichen Einsatz von KI: (1) Management von Künstlicher Intelligenz, (2) Organisation des Betriebs, (3) Rechtliche Gestaltung, (4) Regulierung und Compliance, (5) Lebenszyklus-Management, (6) Management der Technologie-Infrastruktur, sowie (7) Cybersicherheit.Der vorliegende Artikel leitet konkrete erste Schritte an und richtet sich primär an Geschäftsleitungsmitglieder, IT-und Innovationsverantwortliche sowie Projektleiter, welche die neuen Wertversprechen der KI in der betrieblichen Praxis verwirklichen möchten.

Barriers to the Use of Artificial Intelligence in the Product Development – A Survey of Dimensions Involved

“…Regarding AI, Chapter 2.1 already explains its distinction from traditional software projects. A rethinking of the use of AI technology in companies is necessary (van Giffen et al, 2020). The companies surveyed as part of the representative Bitkom survey also cite a lack of personnel resources and data (62% each) and a lack of technical expertise (48%) as barriers to deployment.…”

“…At this point, care must be taken not to create a missing link between top-down and bottom-up perspectives in scientific literature. AI use will broadly affect a wide range of business areas and functions (van Giffen et al, 2020). Initial approaches involving individuals without broad AI expertise in the development of AI applications use AI Cards as an ad hoc means of mitigating the lack of knowledge (Aranda-Muñoz et al, 2022).…”

“…In AI system development, there are fundamental differences to traditional software development. Projects that follow the traditional software development approach typically provide a software application which is produced by writing a source code, with the software development process typically following one of the established process models, such as the V-model (van Giffen et al, 2020). AI processes, on the other hand, involve having to train neural networks and processing a large amount of data.…”

“…AI processes, on the other hand, involve having to train neural networks and processing a large amount of data. The training process is a highly automated and computationally intensive process that requires extreme computing power to develop the AI models (van Giffen et al, 2020). The integration of AI-based applications into products, services, processes, and business models requires both the backing of and anchoring in senior management, and comprehensive AI knowledge in the organization (Kreutzer and Sirrenberg, 2019).…”

“…Examples of currently known models and methods can be found in the literature. van Giffen et al (2020) present the St. Gallen Management Model for the operational use of AI. Kreutzer and Sirrenberg (2019) present a canvas for AI deployment that can facilitate AI deployment by asking questions in seven domains.…”