Regular readers of our blog already know Dr. Monika Oertner: with her study on para-plagiarism, the Konstanz-based writing-pedagogy specialist documented precisely where classical plagiarism software fails. Now she has published something fundamental. Not a test report, but a model for academic core competencies. At akademische-basiskompetenzen.de she presents a rotatable deltahedron with five vertices: reading, writing, thinking, knowledge and ethos. You can grab it with the mouse, turn it, and click on every edge to see how these competencies support one another. Our verdict: it is worth your time. And these thoughts on the nature of learning at school and university naturally play a central role in our work at Mentafy, too.
The model in brief
The idea is as simple as it is consequential: the five core competencies of a degree programme do not stand side by side, but in mutual dependence. Knowledge supports thinking, reading supports writing, and writing in turn supports knowledge. Each of the deltahedron's nine edges carries two explanations, one for each direction — eighteen relations in total. And at each of these relations, it is noted which kind of AI use has the potential to damage it:
- gathering information via AI
- AI text analysis
- AI-generated outlines
- AI text generation, and
- AI text editing.
Oertner explicitly understands the short explanatory texts at the edges of the deltahedron as food for thought; she has deliberately omitted source references there. It is a position paper in the form of a model: pointed, contentious, and productive precisely for that reason.
Why this model convinces us
It answers the question that is chronically neglected in the integrity debate: why, actually? Why is it a loss when students outsource their writing, as long as a decent text comes out at the end?
The model's answer: because the term paper is not a bureaucratic ritual, but the training ground on which all five competencies are applied and developed at the same time. Writing involves reading, thinking, sorting knowledge and taking responsibility for one's own position. Oertner calls this epistemic writing: thinking on paper. Anyone who hands this process over to a machine does not lose a text — they leave the training almost entirely. And because the competencies support one another, it does not stop at one isolated failure. The interplay is, as the subtitle of the site puts it, delicate.
Precisely this line of thinking also underpins our work. The mind is a muscle; a degree programme is its gym. An AI that takes over the training solves tasks, but makes nobody stronger.
The key concept: attributability
One concept from the model deserves particular attention, because it bridges writing pedagogy and integrity checking: attributability (Zuschreibbarkeit). Under "ethos supports writing", Oertner formulates the requirement that every statement in an academic text must be transparent and traceable in its origin. Own and external contributions clearly separated, the provenance of every claim verifiable.
This is nothing less than the theoretical foundation of what we implement technically at Mentafy. Our tools examine exactly this provenance: Semantic Source Search (S³) traces the origin of content even when paraphrasing or translation has erased every verbatim trace. The reference check examines whether the sources cited actually hold up. And the analysis of the writing process can — where such data exists — show whether a text has grown through genuine thinking on paper. Oertner describes the why. We are building the how.
Where we stand — an honest assessment
Full transparency: we are an AI company ourselves. We do not share every sharp edge of the model, and we believe — contrary to what its consistently cautionary tone might suggest — that AI can be a blessing in education: as a tutor that helps with understanding, rather than a ghostwriter that replaces it. We are working on intelligent tutoring systems ourselves.
But on the diagnosis we agree with Oertner completely: wherever competencies are to be acquired and demonstrated, AI must not quietly take over the practice. Learning has to remain learning. Our answer to this is not a prohibition at every edge, but the protection of the spaces in which the training takes place: writing assignments whose authorship can be fairly established. Under our corroboration standard, no finding rests on a single tool, a finding is not a verdict, and in the end a human being makes the decision. That way the unsupervised writing assignment — the training ground of the deltahedron cited here — remains assessable even in the AI era. And what remains assessable does not have to be abolished.
Give it a spin!
The deltahedron is not a text to read, but an instrument: you turn it, click on an edge, and you have a conversation starter for the next faculty discussion, the next teaching conference, the next seminar on good academic practice. That is exactly what it was made for.
It comes with an extensive list of the consequences of AI use in education and society, supported by the very latest sources. And under the "Verwendung" (Use) tab, Oertner offers free teaching materials that make the model usable in the classroom. All of it can be found at akademische-basiskompetenzen.de.
About Dr. Monika Oertner
Dr. Monika Oertner is an author and writing advisor at HTWG Konstanz. Her work on generative AI and academic writing: oertner.net/Publikationen/GKI. The 3D model of academic core competencies: akademische-basiskompetenzen.de.
Further reading on Mentafy
- Para-Plagiarism by AI: Why the Most Expensive Plagiarism Software Misses It — our joint investigation with Dr. Oertner.
- After Adelphi: Why "AI Detection Alone" Is No Longer a Legally Defensible Standard — where the corroboration standard comes from.
No comment yet, add your voice below!