Why What Is Not Said Defines Communication, Medicine, and Democracy
By Angela Maria Carlucci | Michael von Forstner | Gianni Pittella
Image: The Authors
Editorial Note | 31 March 2026
This essay was first published on 24 March 2026 – ahead of the European Parliament’s formal vote on the AI Omnibus (26 March) and the opening of trilogue negotiations.
Published open access – DOI: https://doi.org/10.5281/zenodo.19889108
March 2026 | The age of Generative AI has brought about an ontological rupture. While algorithmic systems excel at processing codified information, they remain fundamentally blind to the meta-levels of silence – the pauses, hesitations, and deliberate omissions that constitute the core of human judgment, clinical diagnostics, and democratic trust. By integrating Strategic Communication, Clinical Medicine, and Democratic Policy, the authors argue that protecting the unquantifiable is a functional necessity for the survival of human-centric institutions.
This article is also available in German | Italian.
Press Review: The Media Debate
QUICK NAVIGATION
The Meta-Levels of Language
Silence in Communication – A Cultural Dimension | Angela Maria Carlucci
Silence in Medicine – A Diagnostic Dimension | Michael von Forstner
What AI Hears – and What It Will Never Understand
Towards a Taxonomy of Silence
Silence and Democracy – A Political Dimension | Gianni Pittella
The Architects of Context
The Meta-Levels of Language
Language is measured. Transcribed. Analyzed. Optimized. In a world of algorithmic communication, what counts as relevant is what is spoken – what can be quantified, processed, and reproduced. But language does not begin with words. It begins with silence. Silence is not the absence of communication. It is its deepest layer – a meta-level that frames, contextualizes, and sometimes contradicts what is said. Those who cannot read silence understand language only halfway.
And yet, in the age of AI, silence is precisely what gets lost. Every system trained on language is trained on words – on what was said, written, or transcribed. What was not said, what was withheld, what hung in the air between two people in a room – this does not exist in the dataset.
This article explores that gap from three distinct and complementary perspectives: from communication and intercultural strategy, where silence is a cultural signal that no algorithm can reliably decode; from medicine and patient safety, where the hesitation before an answer is often more diagnostic than the answer itself; and from politics and democratic theory, where the silence of institutions and citizens is one of the most consequential – and most misread – forms of speech.
Three disciplines. One argument: silence is not a data gap. It is where meaning lives.
Silence in Communication – A Cultural Dimension
– Angela Maria Carlucci
Working at the intersection of strategy, culture, and diverse linguistic landscapes reveals what no algorithm can replicate: the silence that carries meaning.
Effective communication starts long before a post is written or an announcement is shared. It begins with presence, careful listening, and a deep understanding of your audience. In communication, just like in music, if the rhythm is off, the most beautiful melody is useless. C’est le ton qui fait la musique.
In intercultural communication, silence is never neutral. It carries cultural tropes – culturally encoded patterns of meaning that deviate from the expected norm. Where the Code is the rule, the trope is the meaningful exception. In Mediterranean cultures, silence can mean disapproval – or emotional overwhelm, or the deliberate withholding of consent. In Nordic and Scandinavian cultures, the same silence signals reflection, respect, and careful consideration. What reads as patience in Stockholm reads as indifference in Rome. Silence is context-dependent, situational, and irreducibly local. The trope cannot be standardized. The algorithmic Code cannot decode the Trope.
Practical experience in mediating between diverse partners – such as a fast-paced international player and a deeply rooted local institution – demonstrates that language is seldom the primary obstacle. Even with a common vocabulary, projects often stall, revealing that working culture is not found in a dictionary. What created a near-fatal communication vacuum was not a missing word. It was a missing interpretation of silence. For one partner, no news meant chaos or disinterest – silence was read as rejection. For the other, the same silence simply meant: everything is on track, no need to worry.
Silence is not a failure. It is a signal. It marks the boundary between information and trust, between what is said and what is meant. This is the ‘communication vacuum’ – the space where silence is misread across cultural lines, where the absence of a signal becomes the most dangerous signal of all.
In strategic communication, the instinct for the ‘right silence’ is one of the rarest and most valuable competencies. When do you stay silent in a negotiation? In a conflict conversation? In a meeting where a leader speaks – and the silence of those present says more than any contribution? As a Cultural Broker, the real work begins precisely where language ends – not as a translator of words, but as a broker of expectations. In the space between words, the decisive signals form.
AI can generate texts, adapt tonalities, and translate in seconds. But it does not know the history behind the silence. It does not know whether this hesitation signals uncertainty, shame, resistance – or simply reflection. It does not sense when a message is culturally miscalibrated – before the damage is done. It does not know when to speak. And when to stay silent.
AI hears the frequencies. But it does not hear the silence between them. True communication is not only information. It is presence. Judgment. The instinct for the right tone at the right moment – in the right culture. This is not learned from data. It is lived.
Silence in Medicine – A Diagnostic Dimension
– Michael von Forstner
In clinical medicine, silence is an important measure. It is almost as diagnostically relevant as a blood pressure reading or a laboratory result – but it cannot be recorded in an electronic health record, transcribed by ambient listening software, or flagged by an algorithm. It lives in the room. And it disappears the moment the room is no longer attended to by a human being trained to read it.
In the patient-physician dialogue hesitation before an answer is one of the most significant – and most overlooked – moments in clinical communication. When a patient pauses before responding to a sensitive question about pain, mental health, family history, or lifestyle, that pause carries information. Its length, its quality, and its emotional weight tell an experienced clinician something that no transcript can capture. The patient who answers too quickly is also communicating. The one who looks away before speaking. The one whose silence is not the silence of not knowing – but of “not yet being ready to say.”
These are diagnostic parameters that live in the silence between question and answer. They are not anomalies. They are data – of a kind that no current AI system is equipped to process.
The emergence of ambient listening technology in clinical settings raises a fundamental question: What happens to the diagnostic silence when AI is in the room? If used merely as a replacement for human attention, the microphone captures the words but erases the hesitation. The transcript records what was said; it erases what was withheld. And the clinician who is watching a screen – waiting for the AI to suggest a diagnosis – is no longer watching the patient.
However, if integrated correctly, AI offers a transformative potential. By handling the mechanical protocol of the text, it can return the “gift of time” to the physician. An experienced clinician, liberated from the burden of manual documentation, can finally redirect their full sensory attention to the patient. As argued in NEJM AI (Haug & Harrison, 2026), the “human-in-the-loop” principle requires a precise definition. We contend that the true value of the human in this loop is not data entry, but the capacity to remain present in the silence while the machine processes the signal. The goal is an AI that records the text so that the doctor can finally, once again, attend to the context.
These questions are already entering the clinical arena. In June 2026, the theme of silence as a diagnostic meta-level will be introduced into discussions on Patient Centricity and Digital Health at the Royal Society of Medicine – a signal that what AI cannot hear is becoming impossible to ignore.
What AI Hears – and What It Will Never Understand
AI can transcribe, analyze, and translate language; it can detect tonality, identify patterns, and flag anomalies. But for AI, silence is a data gap –an error in the signal, not a signal in itself. This is the fundamental difference between data processing and human communicative intelligence. AI is structurally blind to silence because it cannot inhabit the context. A human being interprets silence through coordinates that an algorithm, by its very nature, cannot comprehend: Who is silent? In what situation? After which sentence? With what expression? In which culture? With what history?
The human-in-the-loop principle is widely recognized but, as highlighted in the recent scientific debate (Haug & Harrison, NEJM AI, 2026), it remains dangerously ill-defined. In the clinical setting, ambient listening is proposed as a transformative solution: a catalyst for humanity capable of freeing the professional from the burden of documentation to restore the “human side” of eye contact.
However, this possibility is not an automatic achievement. The “liberation” of the gaze in ambient listening hides an operational pitfall: if the recovered attention is not intentionally directed toward listening to the unspoken, an ontological rupture occurs. Silence decays into a technical void, excluded from the category of meaning because the machine is structurally incapable of receiving it. In this gap, the risk of technocracy emerges: the algorithm becomes the sole arbiter of what “exists” in the verbal record, effectively expropriating the human being of the power to define the reality of the interaction. The human-in-the-loop principle must therefore be reclaimed as a non-delegable decisional act regarding the significance of silence. Technology opens a space that only human ethics can fill, preventing the reality of human interaction from being mastered by an algorithmic code.
Towards a Taxonomy of Silence
We propose a first classification – provisional, open to debate, and deliberately incomplete. It not as a definitive map but as an invitation to think more rigorously about what we lose when we reduce communication to what is said. Language has always been classified, codified, and taught. Silence has not. Yet silence is as structured as speech – it operates according to its own logic, its own cultural grammar, its own diagnostic weight. Communication theorists have long established that one cannot not communicate. Paul Watzlawick’s foundational axiom – “Man kann nicht nicht kommunizieren”, or the impossibility of not communicating (Pragmatics of Human Communication, 1967) – reminds us that every behavior – including silence, stillness, and absence – carries communicative weight. There is no neutral state. The patient who says nothing, the negotiator who pauses, and the institution that does not respond are all communicating. Silence is not the absence of communication; it is one of its most potent forms.
And yet, for AI, silence remains a blind spot – a gap in the data rather than a signal within it. This is the paradox at the heart of algorithmic communication: a system that processes everything that is said, and nothing that is not.
Diagnostic silence – In medicine, silence is not a gap; it is data. The patient who hesitates before answering a sensitive question communicates something no transcript can capture: the weight of the pause and the distance between the question and the answer. This hesitation, signalling shame, fear, or the difficulty of translating lived experience into clinical language, is often more informative than the answer itself. A physician trained to read this silence holds a diagnostic instrument that no AI currently possesses.
Strategic silence – In negotiation, silence is a move. The deliberate pause after an offer creates pressure without aggression, communicating confidence and opening space for the other side to reconsider. Experienced negotiators know that the first person to speak after an offer often loses the advantage. This is not instinct – it is a learned, culturally calibrated skill entirely invisible to an algorithm analyzing a transcript.
Cultural silence – The same silence carries opposite meanings depending on context. What reads as respect in one culture reads as evasion in another; agreement in one context signals deep disagreement in another. There is no universal grammar of silence, only local codes requiring years of immersion and observation to decode. This is why failures so rarely happen at the level of vocabulary, and so often in the spaces between words.
Institutional silence – Within organizations, collective silence is a frequently misread signal. A quiet meeting may reflect genuine consensus, or it may signal fear, dissent, or the collapse of psychological safety. Leaders who cannot distinguish between these forms make decisions on false premises. What is not said or acknowledged shapes an organization’s culture as powerfully as any official statement.
Political silence – In democracy, silence is both a right and a weapon. The silence of institutions in the face of injustice is a form of speech; the silence of disillusioned citizens is a democratic crisis. Technocracy cannot tolerate the unquantifiable and the ambiguous, yet this irreducibly human silence is precisely what must be protected to preserve the productive ambiguity of democratic deliberation.
Digital silence – In the age of AI, silence is the ultimate blind spot. What the machine does not record simply does not exist in its model of the world. The gap in the data is treated as an absence of meaning, yet it is often the location of the most concentrated truth. The patient’s hesitation or the negotiator’s pause determines whether a conversation succeeds; the machine hears the frequencies, but it never hears the silence between them.
This taxonomy is not exhaustive; it is a beginning. Each of these silences deserves its own study, methodology, and pedagogy. What unites them is this: they are all forms of meaning that exceed the capacity of any algorithmic system to capture.
Silence and Democracy – A Political Dimension
– Gianni Pittella
In an era dominated by an algorithmic obsession with immediate response and constant quantification, we must rediscover the exquisitely political value of silence. If technology aims to fill every void with the efficiency of data, politics must reclaim the right to the interval and the pause: not to defend an indefensible opacity, but to protect the very possibility of reflection. For it is only through reflection that one can achieve mediation.
Democracy is not an uninterrupted flow of information, but a delicate balance of listening, waiting, and judgment. To betray this silence is to yield to a cultural and economic trend – one that is highly influential and dangerous today – which claims that democracy is an obsolete system in irreversible crisis. This thesis argues that we should deliver a final “blow” to representative institutions by over-emphasizing Artificial Intelligence, replacing them with a technocratic government.
We must be clear: this transition from democracy to technocracy is, in fact, an evolution toward autocracy. It is a shift that we must absolutely prevent. The illusion that every human dilemma can and should be resolved by an automated process devoid of political responsibility neutralizes freedom itself. Within the architecture of institutions – particularly European ones – silence is the necessary space for synthesis between divergent visions. If the “Code” becomes the sole rule, we lose the ability to interpret the exception and to govern the complexity of reality.
Politics requires “shadow zones” that protect the integrity of the decision-making process and allow for the internal maturation of consensus. Artificial Intelligence can analyze the frequencies of consensus, but it can never grasp the depth of an institutional silence in international negotiations. The true leader knows how to inhabit silence to draw from it the strength of choice. Sovereignty, in the 21st century, is the power to remain unquantifiable. It resides in the human capacity to decide when to speak and when, instead, to let silence mark the boundary of our freedom. We must recognize that what cannot be measured is not therefore without value. Indeed, it may be the most valuable thing we have. Because true politics is not learned from data. It is lived.
The Architects of Context
Language is far more than what is said. True communicative competence – whether in a doctor, a negotiator, a leader, or a diplomat – lies in the ability to read what is not said; to stay in silence without rushing to fill it; to distinguish the silence of doubt from the silence of reflection, cultural silence from emotional silence, and diagnostic silence from defensive silence. This is the silence of meaning, and it cannot be algorithmized. It cannot be scaled. It cannot be reproduced by any artificial intelligence system, however sophisticated. It is human – deeply, irreducibly human.
But there is a second silence, and it is more dangerous. It is the silence that algorithmic systems impose: the blindness of systems that do not register what they cannot measure, making invisible precisely what most needs to be heard. This is not the silence of meaning; it is the silence of erasure.
Silence is not outside language. It is above it. It is the meta-level that gives language its meaning, its weight, and its humanity. Without it, communication is data. With it, it is life. And perhaps it is precisely there – in defending the silence of meaning against the silence of erasure – that the last territory of our freedom resides.
Angela Maria Carlucci – Strategist and Director of Sintagma. Global Communications, Social Dialogue, and Public Affairs.
Michael von Forstner – Pharmacovigilance Executive. Founder of Mesa Laubela and MedGenie AG. Patient Safety and Clinical Risk Management. Fellow of the RSM.
Gianni Pittella – Physician and Policy Advisor. European and International Policy. Former First Vice-President of the European Parliament.
ABOUT THE AUTHORS
Angela Maria Carlucci is an Italian-Swiss strategist and director of Sintagma. She holds a degree in Foreign Languages and Literatures from the University of Florence, with academic studies in Frankfurt and Toronto. Examining the ontological rupture of generative AI through foundational semiotics, she bridges theoretical research with strategic consultancy for global markets, exploring the human dimensions of AI. Her perspective is informed by extensive experience in global corporate strategy, communications, and Swiss and EU social dialogue.
Michael von Forstner is a pharmacovigilance and patient safety executive, founder of Mesa Laubela and MedGenie AG. He studied biochemistry and medicine in Graz, holds a PhD from ETH Zurich, and an MPH from the London School of Hygiene and Tropical Medicine. A former researcher at UC Berkeley and assistant professor at SLU Uppsala, he is a fellow of the Royal Society of Medicine (RSM), with vast experience at Roche, Boehringer Ingelheim, and Biogen, specializing in digital health. He is currently introducing the theme of silence as a diagnostic meta-level into discussions on Patient Centricity and Digital Health at the RSM.
Gianni Pittella is an Italian politician and physician. He graduated in medicine and surgery and specialized in legal and forensic medicine at the University of Naples Federico II. He was elected Member of the Italian Parliament in 1996 and Member of the European Parliament from 1999 to 2018, serving as First Vice-President of the European Parliament from 2009 to 2014, and as President of the Progressive Alliance of Socialists and Democrats from 2014 to 2018. Following his tenure as a Senator of the Italian Republic (2018–2022), he is active as a senior advisor on European and international policy.
© 2026 | Carlucci | von Forstner | Pittella
All rights reserved. Permission is granted for journalistic reproduction. No unauthorized AI harvesting.
Press Review: The Media Debate
THE WATCHER POST – Le molteplici forme del silenzio, il pamphlet che guarda all’AI per ciò che è –
(The pamphlet that looks at AI for what it is)
PENSIERI MERIDIANI – Le molteplici forme del silenzio (full editorial in Italian)
«The essay produced by Sintagma.com – led by experts Angela Maria Carlucci, Michael von Forstner, and Gianni Pittella – aligns with the Mario Draghi Agenda for the Old Continent, exploring the critical intersections between human silence and Artificial Intelligence.» — Gianfranco Blasi, Author, former Member of the Italian Parliament and of the Budget, Treasury and Planning Committee.
NUOVO GIORNALE NAZIONALE – Le molteplici forme del silenzio – A cura di Angela Maria Carlucci
Parte 1: Comunicazione | Parte 2: Medicina | Parte 3: Politica
LE CRONACHE LUCANE / BASILICATA 24 ORE – Le molteplici forme del silenzio
Parte 1: IA generativa e linguaggio umano | Parte 2: Medicina clinica | Parte 3: Democrazia – una dimensione politica
LE CRONACHE LUCANE – Le molteplici forme del silenzio
MED – Le molteplici forme del silenzio – in prima pagina / Versione PDF
IL BELLO – Tecnologia e democrazia, le molteplici forme del silenzio – in prima pagina