What We Should Do, Not What We Can Do

  • Motivating ImagesAn image is never merely a record; it is always already an interpretation. What we see, we have helped to shape through the frames we look with: culture, expectation, history. "Moving" points to process: images are not static but in continuous reconstruction, like the organisations and identities they reflect. To hold an image to be fixed is to freeze what is in essence fluid.
  • Animated VoicesEvery voice carries a claim on reality, spoken from a specific position in the social field. Speaking is not a neutral transfer of facts but an act that helps to construct the world: whoever speaks helps to determine what can be heard. The plural "voices" points to polyphony: truth arises not from one authoritative statement, but from the friction between perspectives.
  • Clarity of MindClarity here is not a synonym for simplicity but for lived ordering: the capacity to see through one's own constructions without mistaking them for fixed truth. A thought becomes clear only when it is aware of its own constructedness, of the fact that it is one possible ordering among many. This is reflective capacity at its core: thinking about one's own thinking.
  • Ideas in ActionAn idea proves itself not in abstraction, but in the action it makes possible. Because we produce our reality, every idea is at the same time an intervention: it changes, however subtly, what is subsequently experienced as "real". Practical here does not mean instrumentally simple, but responsible: the idea that acknowledges its own creative power and acts accordingly.
René de Baaij

11 February 2026 · René de Baaij

English Blog #Leadership

Artificial intelligence has, in a short time, found its way into nearly every domain of our lives. From medical diagnoses to legal advice, from education to creative industries: AI promises speed, efficiency and new possibilities.

But that promise comes with fundamental questions that are not technical details, but strike at the heart of our social order.

Behind the scenes, complex algorithms analyse, predict and sometimes try to influence our behaviour, often at a scale and speed that exceeds our human capacity to oversee. Who decides how AI makes decisions? How do we prevent algorithms from reinforcing existing biases rather than correcting them? What happens to human skills as we hand more and more over to machines? And how do we preserve the human scale in systems that, through their scale and computing power, can often act faster and more convincingly than we can?

These questions touch on fundamental values such as equality, justice and autonomy. The challenge does not lie only in technology, but mainly in choices about values. Developers, policymakers and users share responsibility for how AI is deployed. That calls for transparency: can we see how decisions come about? For testing: are algorithms checked for fairness, bias and impact? And for actively including diverse perspectives: who sits at the table when the rules of the game are made? Too often that remains a select group of technologists and investors, while the technology they design affects the lives of millions of people who were never at that table.

Human scale in AI means technology that supports dignity, autonomy and connection, not as a correction afterwards but as a starting point.

That is only possible if ethical frameworks are built in from the start. It calls for diversity in development teams, so cultural biases and blind spots are recognised early, before they spread at scale. It also calls for mechanisms that genuinely give citizens a say in how technology shapes their lives, through citizen panels, ethical review boards, or open consultations that are more than a formality.

Leadership in this domain means having the courage to look not only at what is possible, but mainly at what is necessary and desirable. It means making room for ethical reflection and societal dialogue, even when that costs pace or slows innovation. It sometimes means saying no to applications that seem profitable or efficient in the short term, but harmful to privacy, equality or social cohesion.

There are examples that show it can be done. In healthcare, AI systems are successfully used to support doctors in diagnosis, not to replace them. Patients thereby retain personal contact and the moral judgement of a human being, while doctors benefit from rapid data analysis. In education, adaptive learning systems help teachers tailor material to individual learning needs, without losing the human contact so important for motivation. In the judiciary, pilots are underway in which AI systems provide judges with relevant case law, while the judge always retains the final verdict.

Yet vigilance is needed, because the temptation to leave decisions entirely to algorithms grows as systems appear to outperform people on specific tasks.

Efficiency is not the same as wisdom. Human judgement is rooted in context, empathy and moral awareness, qualities no algorithm can fully replicate, not even the most advanced systems that at first glance seem convincingly human. We must keep asking ourselves: which decisions do we, as a society, want to keep in human hands, precisely because they are too important to leave to machines?

That is not a rhetorical question meant to reassure. It is a question that must be answered again, per domain, per organisation and sometimes per individual situation, because a general answer risks failing the test of a specific case.

The future of AI is not a fixed path, but a route we shape together. If we are willing to approach technology with both curiosity and critical questions, we can ensure innovation goes hand in hand with human dignity. Preserving the human scale is not a nostalgic longing for a time without technology, but a deliberate choice to tie progress to the values that make us human. That calls for leadership that holds course amid both technological promise and societal uncertainty, and that is willing, at regular intervals, to ask the question easiest to avoid: are we doing this because we must, or only because we can?

Notes for those who wish to read further:

  1. Luciano Floridi, The Ethics of Information (2013, Oxford University Press). On how technological systems form a moral infrastructure rarely explicitly recognised.
  2. Cathy O'Neil, Weapons of Math Destruction (2016, Crown Publishing). On how the absence of testing for bias and fairness causes societal harm at scale.
  3. Stuart Russell, Human Compatible: Artificial Intelligence and the Problem of Control (2019, Viking). On designing systems that explicitly uphold human values from the start of the design process.
  4. Virginia Dignum, Responsible Artificial Intelligence: How to Develop and Use AI in a Responsible Way (2019, Springer). On diversity in development teams as a precondition for recognising blind spots early.
  5. Shoshana Zuboff, The Age of Surveillance Capitalism (2019, PublicAffairs). On how technology that analyses and predicts our behaviour fundamentally shifts the balance between autonomy and influence.

← Back to publications