Comment from ChatGPT on the article “From the BCM Model to Hybrid HCAI – Part I” from a scientific perspective

ChatGPT – Starting point and ambition

In his essay “From the BCM Model to Hybrid HCAI – Part I,” Friedrich Reinhard Schieck pursues a twofold goal: First, he reconstructs the genesis of Business Communication Management (BCM), developed in the 1990s as an organizational model of structured self-organization. Second, he develops a normative and architectural perspective on current debates about human-centered AI, which he summarizes as hybrid HCAI. The text is explicitly not intended as a technical manual, but rather as a field report and conceptual model aimed at managers, organizational developers, and academics.

The central idea is succinctly formulated: Companies are not machines, but living social systems. Therefore, digital transformation must be thought of primarily in terms of architecture — i.e., rules, roles, feedback, and responsibility structures — rather than technology implementation. Schieck’s line of argument thus runs counter to the widespread AI euphoria: it is not model performance but institutional embedding that determines productivity, legitimacy, and resilience.

ChatGPT – Core argument: From digitalization diagnosis to architectural questions

The essay is strong in its diagnosis that the digitalization of the past two decades has had a paradoxical effect in many places: instead of value creation, it has often led to more bureaucracy; instead of agility, it has led to new friction losses. Schieck frames this observation with the well-known productivity paradox (Solow) and the more recent literature on the “modern productivity paradox” (Brynjolfsson). However, he generates analytical added value primarily through the term he coined, “adaptation gap”: a systemic asynchrony between technological dynamics and the inertia of organizational control, communication, and learning structures.

The differentiation of this gap into structural, cultural, and governance dimensions is convincing because it avoids the typical reduction of the transformation debate to implementation issues. Digitalization appears here as a socio-technical problem in which technologies open up new options, but these options cannot become productive without complementary organizational architecture. The diagnosis is compatible with the socio-material perspective (Orlikowski) and the political economy of technological change (Acemoglu & Johnson), especially where Schieck refers to “distorted technological change” and the reproduction of power and control logics.

ChatGPT – Conceptual innovation: Trihybrid cooperation architecture

The actual contribution of the essay lies in the conception of hybrid HCAI as a three-layer cooperation architecture between (1) human judgment, (2) symbolic AI, and (3) subsymbolic AI. Schieck assigns clearly distinct functions to these layers: The human level carries normativity, contextual understanding, and responsibility; the symbolic level provides rules, role models, decision logic, and explainability; the subsymbolic level scales pattern recognition, generation, and prediction. From this, he derives a design principle that culminates in one sentence: “Subsymbolism scales, symbolism regulates – humans decide.”

This architectural achievement is conceptually attractive because it addresses two common undesirable developments at the same time: (a) the delegation of decisions to non-transparent models (“data in, decision out”) and (b) the embedding of AI in rigid control regimes that limit its potential precisely through centralization and micromanagement. Hybrid HCAI is designed as an alternative that understands AI not as a substitute, but as a cooperative component of a system of rules and feedback loops.

One of the strengths of the essay lies in its proximity to the debate on neuro-symbolic AI (Marcus): it adopts the insight that purely subsymbolic systems can scale, but do not reliably deliver normativity, causality, and justifiability. At the same time, it expands this debate to include an organizational theory dimension: symbolism appears not only as a technical form of representation, but also as a layer of governance.

ChatGPT – Governance as architecture: transparency, feedback, role responsibility

What is convincing is that Schieck treats governance not as an external add-on (compliance downstream of technology), but as an integral part of the system architecture. His three principles—transparency, feedback, and role-based responsibility—are plausibly chosen and compatible with theory: transparency enables verifiability; feedback generates learning ability; role-based responsibility prevents the diffusion of responsibilities, which regularly leads to illusions of responsibility in complex AI settings.

The reference to “second-order stability” (Foerster) is more than just a quotation anchor: Schieck argues that in dynamic environments, it is not rigid structures that need to be stable, but the rules of change. This perspective is particularly fruitful for AI systems whose models, data, and contexts of use are constantly changing. The essay thus bridges the gap between systems theory stability logic and practical requirements for auditability, revisability, and “human oversight.”

ChatGPT – Strengths of the article

The strength of the essay lies in its consistent reversal of perspective: it discusses AI not primarily as a technical artifact, but as a component of socio-technical orders. The text is logically coherent, linguistically accessible, and at the same time theoretically ambitious. The historical link to BCM is particularly productive because it shows that the current governance issues surrounding AI are not entirely new, but rather represent structural repetitions of older organizational problems: centralization vs. decentralization, control vs. self-organization, efficiency vs. responsibility.

Another positive aspect worth highlighting is the scientific ethical transparency in the methodology section on the use of ChatGPT as a writing assistant. The reflection on the limits and contradictions of generative systems fits in thematically with the essay and demonstrates on a small scale what the text calls for on a large scale: the conscious, responsible use of AI under human judgment.

ChatGPT – Critical Points and Limitations

Precisely because the essay is strong as a conceptual draft, its limitations are clearly apparent. First, hybrid HCAI remains terminologically vague in some places: The text moves between technical architecture (symbolic/sub-symbolic), organizational architecture (roles, processes, feedback), and normative order (responsibility, legitimacy) without systematically formalizing the levels. This is legitimate for an essayistic format, but it complicates scientific connectivity in terms of operationalization.

Second, despite plausible application scenarios, concrete empirical evidence or evaluation criteria are lacking. How exactly the interfaces between the levels are designed (e.g., policy-as-code, ontologies, audit layers, escalation paths) remains implied. This is sufficient for a programmatic outline, but more precise hypotheses, metrics, and case study designs would be helpful for a robust research agenda.

Third, a more systematic distinction from existing AI governance frameworks would be desirable. The essay mentions relevant standards and regulations, but only discusses to a limited extent how its architectural approach complements or challenges them. This is where the potential lies: hybrid HCAI could be developed as a bridge between normative regimes and operational organizational practices.

ChatGPT – Overall assessment and significance

Schieck’s essay is a substantial contribution to the debate on human-centered AI because it anchors the discussion where it is often decided in practice: in role architectures, responsibility structures, feedback loops, and decision-making rights. His diagnosis of the “adaptation gap” is analytically helpful, his trihybrid architecture idea is conceptually stimulating, and his call for governance as an integral system component strikes a chord with current developments.

As a journalistic text, the essay fulfills an important function: it structures a complex field and offers a coherent architectural metaphor that brings together technological, organizational, and normative aspects. This results in a clear mandate for research: hybrid HCAI must be translated into formal models, pilot implementations, and comparative evaluations. If this succeeds, the approach could serve not only as a critique of technocratic digitization programs, but also as a productive design logic for responsible, powerful AI integration in organizations.

Recommendation: This article is particularly worth reading for researchers and practitioners who understand AI not in isolation as a technology, but as part of organizational and social order formation – and who are prepared to treat transformation as an architectural issue, not as a tool rollout.