Professional rationales and interpretations of autonomy

Although the two sessions draw on different professional contexts, they remarkably often define professional autonomy in terms of similar core principles. In both programmes, professional added value is shifting away from mere execution and towards understanding, assessing, monitoring, interpreting and taking responsibility within AI-integrated work processes.

Both sessions make it clear that, in a context of structural AI integration, professional autonomy no longer means that professionals carry out everything themselves, but rather that they retain sufficient understanding to critically assess output, steer processes and intervene in a targeted manner when systems fail or fall short. Understanding, control and responsibility remain central professional expectations in both domains.

At the same time, the sessions demonstrate that the nature of this autonomy remains heavily dependent on the professional context in which AI is used. The core of autonomy is therefore comparable, but the objects to which that autonomy relates differ fundamentally between the two degree programmes.

Within CMM, autonomy is primarily linked to the creation of meaning, creative direction and relational alignment. The professional must be able to interpret a brief beyond the literal request, make creative choices from their own perspective and position output within a broader contextual and relational framework. The concern here is less about efficiency or technical functionality, but primarily about preserving creative individuality, a distinctive style and substantive distinction in a context of heavily AI-generated output. The term ‘one-size-fits-all’, which recurred repeatedly during the session, makes this concern particularly evident.

A key insight from the CMM session is that professional added value lies less and less in the production itself, but is shifting increasingly towards defining and monitoring the framework within which AI generates content. Wouter (Yungo), for example, describes a scenario in which AI produces millions of variations of content, whilst the professional remains responsible for the ‘overarching concept’ from which that output stems. It is precisely this defining, monitoring and justifying of creative frameworks that emerges within CMM as the non-delegable core of professional autonomy.

Within TI, professional autonomy takes on a more technical and systemic character. The professional must understand how systems function, be able to analyse generated code, detect errors and justify architectural choices in substantive terms. Control, reliability and responsibility form the central professional standards in this regard. The statement “The moment something goes into production, AI ceases to exist” illustrates this logic based on a participant’s concrete experience. However, the statement should not be interpreted literally to mean that AI cannot play a role in production or deployment processes. Above all, it highlights that, in critical production situations, responsibility cannot be transferred to AI and that professionals must retain sufficient understanding of the system to assess output, diagnose problems and intervene themselves where necessary.

The TI session also adds two important dimensions to professional autonomy: ownership and meta-control. Ownership refers to the willingness to take responsibility for the end result, even when parts of the solution were generated by AI. Meta-control refers to situations in which AI systems can increasingly be deployed to monitor or evaluate other AI outputs as well, thereby shifting the role of developers towards monitoring the control mechanisms themselves. Both concepts demonstrate how professional autonomy within IT is becoming increasingly abstract and systemic.

From the dispositional perspective introduced in this synthesis, the concept of ownership in particular takes on an additional meaning. After all, the sessions suggest that professional responsibility does not rest solely on the ability to make correct decisions, but also on the willingness to continue to take responsibility for decisions and outcomes in which AI played an active role. Ownership can therefore be understood not only as a professional competence, but also as a professional disposition that helps determine how professionals actually shape their autonomy in AI-integrated professional contexts.

Despite these domain-specific differences, both sessions ultimately reveal a similar trend. In both contexts, professional autonomy is shifting from execution to assessment, interpretation, monitoring and taking responsibility. At the same time, the sessions suggest that this professional autonomy is not based solely on knowledge and skills, but also on professional dispositions that determine whether professionals actually exercise their judgement, control and responsibility when AI plays an active role. The specific implementation differs between the two programmes, but the underlying logic remains strikingly similar.

This finding has direct implications for curriculum development. Generic AI competences that focus exclusively on the operational use of AI fall short. Programmes would do well not only to make explicit which knowledge, skills and forms of professional autonomy remain essential within their field, but also which professional dispositions students need to develop in order to apply these competences critically, responsibly and purposefully in AI-integrated professional contexts.