Applications Change, Data Remains the Same


The concept of humanist superintelligence—originally termed „Humanist Superintelligence“ in English—is a cognitive initiative from Silicon Valley that is being promoted by Microsoft’s AI division (MAI) as a new strategic and ethical framework for the development of artificial intelligence.
Behind this lies the ambitious goal of creating the world’s most capable and advanced AI systems, which at the same time remain fully controllable, ethically grounded, and firmly dedicated to serving humanity. Microsoft defines this humanistic foundation through soft, almost philosophical-sounding core values such as kindness, trust, quality, simplicity, security, and continuous evaluation.
For existing SAP customers, this narrative calls for critical analysis, as the gap between ethical IT marketing and the harsh business reality of ERP could hardly be wider. Large language models have already proven unsuitable for ERP because LLMs rely solely on language and statistics. Wisely, SAP CEO Christian Klein has acquired the German startup Prior Labs, where AI systems are developed based on numerical values—because semantic friendliness has no place in an Autonomous Enterprise!
When AI developers assert that the nature of human collaboration defines what is built, and that humanity is at the center of every algorithmic decision, they are deliberately attempting to allay the fundamental societal distrust of uncontrollable black-box systems. Interestingly, the promoted development principle of „kindness“ also appears in socioeconomic discourses, such as in Oona Horx Strathern’s concept of the „Kindness Economy,“ which presents a humanistic alternative to ruthless capitalism and defines a kind, human-centered economy as a decisive competitive advantage.
However, if an existing SAP customer applies this lofty vision to an S/4 transformation and the “era of the Autonomous Enterprise” proclaimed by SAP CEO Christian Klein, Microsoft’s humanistic superintelligence quickly reveals itself to be a necessary, but deceptive construct designed to win over a nervous base of IT users. The challenge does not seem to lie in user-friendliness, but rather in the volatility of agentic AI and the orchestration of ERP data. Algorithms are changeable; data is the foundation of ERP. ERP orchestration is based on data—Master Data Management!
Existing SAP customers should be aware that even an ethically programmed, humanistic AI is based on probabilistic calculations (large language models), which are mathematically bound to be subject to structural hallucinations.
An agent-based AI system that autonomously manages global supply chains or initiates sensitive financial transactions behind the scenes in S/4 HANA cannot be protected from factually incorrect conclusions by mere philosophical guidelines such as „trust“ and „security.“ Microsoft AI’s claim that the systems will remain controllable under all circumstances overlooks the proven information-theoretical limits established by Kurt Gödel and Alan Turing, which irrefutably demonstrate that a perfect, error-free truth machine is simply impossible.
For this reason, the humanistic approach for existing SAP customers is less a guaranteed feature of the software they purchase and more of an imperative, ongoing organizational mission for their own company. Increasing automation through systems such as SAP Joule or external Microsoft models does not render human skills—such as empathy, ethical decision-making, and building trust—obsolete; on the contrary, it elevates them to the most important assets in the workplace of the future.
In the symbiosis of human and artificial intelligence, humans no longer function merely as executors of transaction codes, but assume the indispensable role of a „human in the loop“ or even „human above the loop,“ who is responsible for final decisions, sets ethical boundaries, and critically validates the AI’s results. Any ERP architect who blindly trusts the promises of a humanistic superintelligence and grants it operational control over their ERP system without a rigorous human oversight mechanism is acting with a high degree of negligence.



