The Theory of Complexity and "No Free Lunch" Theorem in Business Artificial Intelligence
Résumé
An increasing number of business applications are using a variety of approaches of Artificial Intelligence (AI) in their ultimate objective of extracting the Business Value. The Business Value originates from market trends, firms competitiveness, customer behaviours, decision capabilities, and so forth. One main sub-field of AI that is widely used mainly in predictive and proactive Business Values is ‘Machine Learning' (ML). However, MA algorithms are plentiful, and most of them necessitate specific Data, particular applicability features, and deep analysis of the Business context in order to learn and then map the model. Because of these considerations, some industrial applications of AI and particularly ML algorithms in enterprises fail. Consequently, the question we address in this paper is “Is there No-Free-Launch Theorem for Artificial Intelligence and Machine Learning applicability in Business?” A focus is put on the Theory of Complexity (TOC) and the “No-Free-Lunch-Theorem” (NFLT) to understand the trajectory of research related to Artificial Intelligence applications in Business. The proposed study of complexity in the applicability of such techniques -characteristics, needs, constraints, and limits- with regard to the business involvedness is suitable for both industrials and academics.