Core competencies and desired level of independence
The cross-case synthesis shows that the desired professional autonomy is interpreted in a domain-specific manner. Although both domains experience similar tensions surrounding AI integration, the underlying professional rationales differ considerably. Curriculum development should therefore not merely formulate AI competences in generic terms, but should always link them to the specific forms of professional judgement, responsibility and substantive understanding that are central to the domain.
Within CMM, the emphasis is more strongly on meaning-making, creative direction, contextual sensitivity, relational alignment and substantive positioning. Within TI, the focus shifts more towards system understanding, architectural reasoning, debugging, analysis of complex dependencies, monitoring performance and security, and ownership within technical environments.
At the same time, both sessions demonstrate that professional autonomy within AI-integrated contexts is not determined solely by the ability to use AI tools. Students must not only develop operational AI literacy, but also learn how in-depth understanding, active AI-steering and professional judgement are interrelated within complex work processes.
From a dispositional perspective, this means that curriculum development should not focus exclusively on which competences students need to acquire, but also on how study programmes can create learning environments in which students learn when, why and with what level of responsibility they apply these competences within AI-integrated professional practices.
The matrices below therefore do not constitute definitive competence frameworks, but rather an initial translation of the scenario typologies into curriculum development within WP3. They highlight which forms of autonomy study programmes aim to strengthen, which risk areas arise in the process, and which curriculum choices are necessary to avoid dependent AI use and superficial competence.
CMM – initial translation into curriculum development
| Professional task | Desired autonomy | AI role | Risk area | WP3 implication |
|---|---|---|---|---|
| Briefing interpretation | Independently interpreting implicit expectations, context and meaning | AI supports structuring | Superficial interpretation and loss of context | Learning activities focusing on context analysis, interpretation and justification |
| Creative concept development | Independently determining direction, positioning and creative signature | AI generates inspiration, variations and prototypes | Generic output and loss of substantive distinctiveness (one-size-fits-all) | Assessment of reasoning, positioning and substantive choices |
| Evaluating AI output | Assessing output in terms of content and actively refining it | AI supports production and iteration | Convincing output without in-depth evaluation | Critical analysis and reconstruction of AI output |
| Monitoring brand consistency | Monitoring content coherence and positioning | AI generates scalable content variations | Reactive correction and loss of coherence | Exercises focusing on active steering, quality control and iterative evaluation |
| Client communication | Justifying professional choices in terms of both relationships and content | AI supports preparation and structuring | Loss of human alignment and professional judgement | Oral defence, stakeholder discussions and reflection |
TI – initial translation into curriculum development
| Professional task | Desired autonomy | Role of AI | Risk area | WP3 implication |
|---|---|---|---|---|
| Requirements analysis | Independently analysing problems, context and dependencies | AI supports structuring and documentation | Analysis without in-depth understanding of the system | Cases involving interpretation, analysis and problem definition |
| Architectural choices | Being able to justify and structure systems in terms of content | AI suggests solutions and patterns | Choices made without insight into implications or dependencies | Justification of design choices and system reasoning |
| Code development | Understanding, adapting and evaluating generated code in terms of its substance | AI generates boilerplate, suggestions and implementations | Working code without an underlying understanding | Code review, reconstruction and debugging |
| Debugging | Analysing and resolving problems independently | AI supports fault detection and analysis | Reliance on AI diagnoses and reactive problem-solving | Problem-driven learning and analysis of errors |
| Production ownership | Taking responsibility for quality, security and maintainability | AI supports the development process | Diffuse responsibility and illusory control | Assessment of ownership, motivation and professional accountability |