Axis 2: Steering – from low steering to high steering

This axis describes the extent to which professionals actively and purposefully control AI systems during the work process.

At one end of the spectrum is low steering, where professionals mainly use AI via the system’s default settings, without targeted control or active adjustment. Prompting remains limited or vague, and output is largely accepted as the system generates it. Professionals respond little or not at all to undesirable, incorrect or superficial output and make no, or only limited, active adjustments to the process.

At the other end of the spectrum is high steering. Professionals use AI in a targeted manner within clear professional frameworks and provide explicit direction to the process in advance through targeted instructions, context, objectives and expectations. They then actively refine the AI output by iteratively refining prompts, formulating additional instructions and responding to shortcomings or deviations in the generated output. High steering presupposes sufficient AI literacy to be able to guide AI systems consciously and purposefully.

Both sessions clearly position desirable professional conduct on the side of high steering, whilst at the same time highlighting how low steering can easily arise in practice. AI systems are becoming increasingly accessible and can be used with minimal technical knowledge, enabling users to obtain useful output quickly without consciously or purposefully guiding the system. At the same time, the increasing complexity and rapid evolution of AI systems heighten the challenge of building up and maintaining sufficient AI literacy. As a result, there is a risk that AI use will evolve into a form of passive reliance, whereby professionals follow the output rather than actively guiding the process. Precisely because AI systems do not always generate consistent or fully predictable output, participants emphasise the importance of active steering, interpretation and fine-tuning by the professional.