Machine-Readable Humanness: Ai, English Abitur, and the Future of Learning in Berlin

---

Deutsch

AI-assisted tutoring system

Thinking With Ai Without Losing Ourselves

That argument stayed with me – not because I agreed with every sentence, but because it touched on something I have been living from the inside while building Sybille, my teacher-built AI-assisted tutoring system for advanced English learners preparing for the Berlin/Brandenburg Abitur (Sybille.Tech).

The real educational challenge of the AI age is not access to answers. It is maintaining human clarity, ownership, and self-regulation while thinking with machines. AI does not merely change productivity. It changes the conditions under which human beings think, write, decide, reflect, and understand themselves. Therefore, the future of education is not simply Ai adoption. It is the design of learning environments that preserve ownership, reflection, context-awareness, and human judgement while working with increasingly powerful systems.

The False Question: Will Ai Replace Thinking in Education?

Much of the public debate around Ai still feels shallow. Parents ask whether Ai will replace essays. Students ask whether degrees will still matter. Teachers ask whether writing itself is dying.

These are understandable questions. I understand the fear. I have even felt some of its social consequences personally. I recently lost a friendship, because I chose to celebrate and continue my work with AI. I will neither dramatize, nor put specific names to the scrutiny and device of mass public media, but the ‘sunken loss’ stands as a reminder of how emotionally charged this subject has become. For some people, Ai is not just a tool. It is a moral line. And I suspect that charge does not go away – it may intensify as the technology deepens. I carry it with me as I write.

Regardless of the fact that when the machine got better at playing chess, we returned to being interested in people playing chess, and Ai art still cannot speak of an ‘artists journey’. If human beings value journey, intention, struggle, authorship, and collective meaning making, not merely optimized output, then perhaps good art doesn’t have much to fear – unless the intent is purely capitalistic. Some say that once the intent is purely capitalistic, then the intent is no longer of art. An aspiring museum director once said to me “Kitsch ist das Gegenteil von Kunst” – Kitch is the opposite of art – and though she retracted the statement shortly afterwards – it is also something that I carry with me as I write.

Still, beneath the fear sits a better question:

What happens to human thinking when answers become permanently available? The real crisis is not that machines can answer questions. It is that human beings may slowly stop recognising themselves inside the thinking process. That is the threat I take seriously. Not the machine itself, but the possibility that we may become invisible to ourselves while using it.

From Prompt Engineering to Context Engineering

The article says that Ai may be imagined as a ‘huge, tireless brain’ full of answers, but that the human task remains knowing what to ask, under which constraints, and what to look for in the answer. It also states that the quality of the answer is governed by the quality of asking.

That speaks directly to something I have noticed over the past months: we have moved from prompt engineering toward context engineering.

The real skill is no longer merely typing a clever command. It is constructing a meaningful environment around the machine: context, constraints, history, intention, tone, purpose, audience, and desired outcome. That may sound new because the technology is fairly new, but the deeper act is not new at all. Law, politics, literature, rhetoric, photography, cross-cultural communication – all of them depend on context. We have always shaped meaning through framing. Ai simply exposes this old truth in a new and sometimes uncomfortable or downright overwhelming way.

This is why language is not becoming less important. It may be becoming more important.

For years, many people assumed technology would reduce the value of language. Yet now, if we want useful Ai outputs, we need better language, sharper framing, clearer intention, and more precise communication. Take the evolution of language around image generation for example. At its best – the quality of outcomes often depends on the traditional vocabulary of photography: angle, lens, lighting, depth of field, contrast, composition, mood. The same applies to any imagery or video generation attempt, primarily because both processes, whether human, or machine generated – are informed by the same body of historical knowledge. The machine may have access to a vaster repertoire of sources, but the human still curates meaning through lived experience, preference, intention, and taste.

The wordsmith is not finished. The wordsmith is threatened on some fronts, yes – but also suddenly central again in different ways.  

Machine-Readable Humanness and the Future of Communication

This leads me to a phrase I find myself returning to ‘machine-readable humanness’. Modern communication increasingly involves translating deeply human intentions into forms that systems can meaningfully interpret without flattening the humanness inside them.

In my own work building Sybille, I have often faced this tension. To what extent do I adjust my natural language so that the builder agent understands me efficiently? At what point does that adjustment become a loss of my own voice? And when the machine speaks back, how much am I being guided, narrowed, encouraged, redirected, or subtly reshaped?

It feels like maintaining a bridge between two riverbanks. On one side is human richness: intention, ambiguity, emotion, memory, humour, cultural background, projection, imagination. On the other side is the machine-readable system: structured input, repeatable logic, constraints, labels, states, tasks, patterns. The bridge must hold. If it collapses, either the machine cannot understand us, or we begin to speak only in the reduced language the machine prefers. The machine has even begun – to in moments refer to my natural way of writing as ‘prose’ – as though style were a deviation from the norm, or ‘inefficient speak’ full of unnecessary adjectives, fluff and embellishments. This tension resonates and at times appears to threaten to erase yet a second bridge – the bridge to myself – because as naturally subjective as it might be, this is the human language of my inner dialogue.

This is why I see the Ai age as a new kind of cross-cultural competence. First, we must become clearer about our own contexts as human beings: our intentions, drives, motivations, values, and interests. Then we have to learn to mediate those contexts into system-readable form without losing fluency in our own human medium. In simpler terms: we need to know ourselves well enough not to disappear into the tool.

Ownership, Reflection, and Responsible Ai Learning with Sybille

---

---

English Abitur, Nachhilfe, and the Human Side of Learning

That matters because the educational challenge is no longer merely producing work. It is remaining visible to yourself inside the work – inside a changing world – while remaining mindful of the merits of tradition.

The inception of Sybille as a character took first form in 2020. As an Ai entity – building started just under 24 Months ago – the first intent being to act as a teacher’s assistant. Sybille’s role has naturally unfolded into what it is now, and over the past 4 months, I have spent a great deal of time constituting philosophies and principles – on reflection and transfer. These philosophies and principles draw in theory and almost 20 years of hands-on teaching experience gathered in a multitude of learning environments. The real implementation and execution of these – now serve ownership – in that Sybille doesn’t only help learners simply to polish the surface, but also to more deeply reflect on the process of arriving at the best results, to explain, defend, revise, present and own the work – and to train replicating exam -aligned performance.

These principles include what I call the Preflight-checks: reflective pauses before work is submitted or presented. The purpose is not merely to improve the final result, but to encourage learners to remain intellectually present within the process itself — aware of what they understand, what they intend, and what genuinely belongs to their own thinking.

That is not an anti-Ai position. It is responsible Ai practice.

Sybille facilitates Personal Learning Plans. A student may need support with false friends, a grammar pattern, register, presentation confidence, or exam writing rhythm. But the purpose is not to outsource growth. The purpose is to build a learning path in which the student remains active, reflective, and accountable.

In summary, Sybille is specialised, not generic, and the pillars upon which she is built include ownership, reflection, concept questioning, responsible AI, and adaptable personal learning plans, age and cultural safeguards, exam-alignment and therefore – pedagogical-integrity. Those are not decorative words for me. They are the educational response to a world in which producing an answer is becoming easier than knowing what the answer means.

Do Not Romanticise Struggle – But Do Not Remove It Either

The article by Dhawan and Raj Sinha also argues that when digital tools can generate comprehensive answers quickly, the patience and discomfort of working out an answer for oneself become more valuable.

I agree – partly.

There is educational value in discomfort. I can say this with relative confidence as a teacher and as a learner, however I would perhaps modify this by saying that the mere act of purposely occupying oneself with the material at hand can also suffice. And though I tend to think that the learning experience can be engaging, fun and fulfilling, there are sides of me that at times embrace the narrative that discomfort is might indeed be the quickest label we attach to the experience in challenging moments. Coincidentally, I am going to the gym (as I do regularly) in a few minutes; I understand the logic of resistance. If the body never receives a signal to adapt, it does not grow. The same is often true of the mind.

But we should not over-romanticise struggle. This holds true to our training metaphor too – as overtraining or ego-lifting can have detrimental cascading effects on the physiology – gains in strength and size are actually realized during periods of rest and recovery, similar to how learning insights sometimes emerge in moments of quiet retreat from the overly structured intensity – during reflection and re-collection.

This oversimplification it introduces a nuanced distinction between: meaningful resistance, and fetishized suffering.

To speak of other metaphors, starting with farming – none of us wants to plough the soil with our bare hands for the authenticity of it.

Not all friction is valuable. The goal is not artificial suffering. The goal is not inefficiency dressed up as virtue. Human beings have always adopted tools: writing, books, sewing machines, cameras, calculators, motor vehicles, computers, smartphones.

When I began photography, some people still treated digital photography as less ‘real’ than analogue photography – portrait painters of the ages needed time to accept photography as art. Later, I found myself laughing at iPhone photography – until I myself began using my phone to photograph, record, and film – while still preserving the nostalgic affections for my old manual camera lenses. The meaning of the stories behind planning a photographic shot with less than convenient tools is not at all lost – though the context it represents is written on another canvas. Technologies change practice. That does not mean nothing is lost. Something is often lost. But something is also gained.

The real question is not whether tools should exist. The real question is which human capacities must remain alive while we use them. For me, those capacities are judgement, discernment, reflection, ownership, intentionality, and communication.

This is where I hold two things simultaneously, and I do not know how they finally settle: Ai may yet make us invisible to ourselves – and it may also return us to ourselves by contrast. I mean both. I cannot resolve them. Perhaps humble acknowledgement of that tension – that irresolution is itself the most honest thing I can do. Creativity is not merely optimisation. Human value will not lie in competing with machine perfection. It may lie in understanding the meaning, ethics, intention, and emotional reality behind the perfectly imperfect human experience.

That is why the future does not belong simply to students who use AI. It belongs to students who can think with it, question it, resist it, shape it, and still recognise themselves in the result.

Students do not only need to think beyond the machine. They need learning environments that help them remain intellectually visible to themselves while using the machine.

That, for me, is where education must go next.

If you are interested in reflective, exam-aligned English Abitur support in Berlin/Brandenburg, you can learn more about Sybille at Sybille.Tech or via TrainingTree.de.

---

Acknowledgement : A recent published article – AI can give answers, but the future belongs to students who can think beyond the machine

The quality of the answer is, in the end, governed by the quality of asking. For curious minds, that is an ever-present opportunity: To expand learning, go deeper across topics, and follow an idea further than was ever possible before

Footnote / descriptor: Sybille is a teacher-built, exam-aligned AI-assisted tutoring system.
Teacher-built, exam-aligned tutoring, purpose-built for structured feedback, safe boundaries, and student ownership.

iAntonio Media – Living the Language: Independent Podcast – @iAntonio_media >> (klassischer RSS-Audio-Podcast)

References:
AI can give answers, but the future belongs to students who can think beyond the machine – Written by: Ashish Dhawan, Pramath Raj Sinha – as posted as at 13.05.2026.
Beyond the Label: Learning Design in an AI Age – Written by: Patterson .
OECD (2026). OECD Digital Education Outlook 2026: The State of Digital Education. OECD.
Kulik, J. A., & Fletcher, J. D. (2016). Effectiveness of intelligent tutoring systems: A meta-analytic review. Review of Educational Research, 86(1), 42–78. (Accessible summary/record.)
VanLehn, K. (2011). The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems. Educational Psychologist, 46(4), 197–221. (Author-hosted PDF.)
TrainingTree article (German)
SPIEGEL Magazin (June 2026). “Viel zu euphorisch oder viel zu kritisch.” Interview with Prof. Dr. Ute Schmid. (Source available as a photograph provided to the author, no public link included.)