The Human Professions
Education, care and social work in an AI-mediated world
The work that cannot be automated
Some professions are built on knowledge, while others are built on relationships. Education, care and social work clearly belong to the latter.
They do not operate only through information, but through presence. Their value is not limited to decisions or outcomes, but is embedded in the interaction itself—in how something is said, how someone is seen and how a situation is understood over time. This makes them fundamentally different from domains that can be reduced to analysis alone.
Artificial intelligence is becoming increasingly capable of processing information, but it does not participate in relationships. And that distinction matters more than it first appears.
The promise of assistance
At the same time, artificial intelligence is already entering these professions in tangible ways. It supports diagnosis, suggests interventions, identifies behavioural patterns and reduces administrative burden. In education, it helps personalise learning pathways and generate feedback, while in social work it assists in structuring complex cases and detecting early warning signals.
These are not marginal improvements. They are meaningful, because they create space—time that can be reallocated and attention that can be redirected. In that sense, AI does not inherently threaten human professions. On the contrary, it can strengthen them.
Yet support is never neutral. The moment a system begins to assist, it also begins to shape.
The shift that follows
Every form of assistance subtly reshapes the activity it supports. When systems structure information, suggest actions and guide decisions, they do more than increase efficiency. They influence what professionals notice, what they prioritise and ultimately what they act upon.
Over time, this changes behaviour.
The question is therefore no longer whether AI will be used—because it already is—but how far it is allowed to shape the interaction itself. And it is precisely in that interaction that these professions find their meaning.
Efficiency versus presence
This is where the tension becomes unavoidable.
Artificial intelligence is designed to optimise for efficiency. It reduces time, simplifies complexity and standardises processes. From a system perspective, this is desirable.
But human professions are not only systems; they are encounters.
A teacher may recognise confusion before a student is able to articulate it. A nurse may sense distress that is not captured in measurable indicators. A social worker may understand a situation that does not fit within predefined categories. These are not inefficiencies to be removed—they are the profession itself.
Yet within system logic, they are easily reframed as costs. Time spent listening becomes delay, attention becomes inefficiency and presence becomes something to optimise away.
This is where the real risk emerges. It is not that professionals are replaced by machines, but that they are gradually pushed to behave like them.
Data versus context
A second tension lies in the way reality is interpreted.
Artificial intelligence operates through data. It identifies patterns, correlations and probabilities, and can signal risks or suggest likely outcomes based on historical information. This can be valuable, particularly in complex environments.
But human situations are not only data points.
They are contextual, shaped by history, environment and emotion. A student’s behaviour cannot be understood without considering the conditions in which it occurs. A patient is not simply a collection of indicators, but a person in a specific moment. A social case is not a file, but a narrative that unfolds over time.
Context cannot be fully captured. It must be interpreted, and interpretation requires judgment.
System versus individual
From this follows a third tension.
Artificial intelligence operates at the level of the system. It scales, standardises and applies general patterns across large populations. Its strength lies in consistency.
Human professions, by contrast, operate at the level of the individual. They deal with exceptions, with deviations and with situations that do not conform to expected patterns. They require flexibility and, at times, the willingness to move beyond the system altogether.
As systems become more optimised, individuals increasingly appear as outliers. And it is precisely in those outliers that professional judgment becomes essential.
The visibility of the human
This leads to a paradox.
As artificial intelligence becomes more capable, the human role does not diminish. Instead, it becomes more visible—particularly in moments where systems reach their limits. Not in routine cases, but in situations of uncertainty. Not where processes run smoothly, but where they break down. Not in efficiency, but in meaning.
The more capable AI becomes, the more visible human presence becomes.
The risk of displacement
There is, however, a structural risk that should not be underestimated.
When efficiency becomes the dominant metric, everything that cannot be easily measured begins to lose value. Relational work becomes secondary, and time spent with people must increasingly be justified in system terms.
In such an environment, professionals are no longer guided primarily by judgment, but by outputs. They are no longer asked to interpret, but to execute.
The profession does not disappear. It shifts. Not by being replaced, but by being hollowed out from within.
What cannot be delegated
At the core of these professions lies something that cannot be transferred to any system.
It is the responsibility not only for decisions, but for people. It is the ability to respond to what is not immediately visible, and the willingness to remain present in situations that are uncertain, complex or unresolved.
Artificial intelligence can support these processes, but it cannot inhabit them. It does not build trust, it does not carry moral responsibility and it does not stand in relation to another human being.
Implications for education
This has direct implications for how future professionals are educated.
If AI increasingly takes over analytical and administrative tasks, then education must invest more—rather than less—in what remains human. Students must learn to engage with ambiguity, interpret context, communicate with sensitivity and make decisions under uncertainty.
These are often described as soft skills. In reality, they are the core capabilities of these professions.
A different kind of expertise
Expertise in human professions is not only cognitive; it is relational. It develops through interaction, through experience and through reflection over time. It depends on the ability to connect knowledge with context, and action with responsibility.
This form of expertise cannot be automated, nor can it be accelerated. It requires time, exposure and engagement.
Europe’s perspective
Within Europe, these professions are closely linked to broader values such as human dignity, public responsibility and social cohesion. Artificial intelligence enters this context not as a neutral tool, but as a force that can reshape how these values are translated into practice.
The question is therefore not whether AI should be used, but whether its use strengthens or weakens the human foundations on which these professions depend.
What remains
What remains, then, is not everything—but something essential.
The ability to be present, to interpret what is not said, to respond to what does not fit, and to take responsibility where systems cannot. These professions are not defined by what they produce, but by how they relate.
Final line
Artificial intelligence can support human professions, but it cannot replace the human relationship at their core.
This article is part of the series The University After AI, published in the Culture & Education section of Altair Media.
Photo by National Cancer Institute / Unsplash
