Published: 12:38 AM, 09 July 2026
Artificial Intelligence, Music and Human Emotion: Gadi Sassoon’s Unusual Performance in Geneva
Author AHM Bazlur Rahman with Shoeb Chowdhury former Senior Vice President of BASIS (2006-2007), one of the iconic figures of adaptation of technology in Bangladesh. -AA
A. H. M. Bazlur Rahman
A United Nations conference hall is usually a place of words. Ambassadors speak. Policymakers warn. Researchers present evidence. Representatives of major technology companies describe possibilities. Civil society asks questions about accountability.
Will AI expand human creativity, or will it gradually begin to exercise power over our feelings, preferences and intentions? At the United Nations’ first Global Dialogue on AI Governance, Gadi Sassoon’s performance brought that uncomfortable question to the stage through art.
But what happens when music suddenly enters the room?
At the opening of the United Nations’ first Global Dialogue on AI Governance in Geneva, the Milan-based composer and transmedia artist Gadi Sassoon began to perform. For a few moments, the debate over artificial intelligence seemed to leave the language of policy papers and enter the more uncertain territory of human feeling.
United Nations video footage and its official shot list captured Sassoon performing before an audience that included António Guterres, the United Nations secretary-general; Doreen Bogdan-Martin, secretary-general of the International Telecommunication Union; and Amandeep Singh Gill, the United Nations secretary-general’s envoy on technology.
Sassoon was not giving a speech.
Yet his music, bodily gestures, digital responses and creative interaction with artificial intelligence appeared to raise a question:
Will AI become a partner in human creativity, or could it one day become an interpreter of our emotions, an influence upon them and, ultimately, a force capable of shaping them?
This is no longer merely a question about art.
It is a question of AI governance.
When Music Enters the Policy Debate
The United Nations’ first Global Dialogue on AI Governance is taking place at a moment when the global debate over artificial intelligence is changing rapidly.
The question is no longer simply what AI can do.
The larger question, in the language of the United Nations, is whether the world will collectively govern this transformation or whether technology will increasingly govern us.
The Dialogue has been organized around four broad areas: the social, economic, ethical, cultural, linguistic and technical implications of AI; bridging AI divides; building safe, secure and trustworthy AI; and respecting human rights through transparency, accountability and robust human oversight.
These issues are normally discussed in the language of policy and technology.
Sassoon seemed to introduce a fifth language: emotion.
Classically trained, Sassoon is difficult to place within the conventional boundaries of electronic music. His work sits at the intersection of orchestral composition, live performance, installation art and agentic AI. In his AI for Good biography, the International Telecommunication Union describes him as a composer and transmedia artist working across precisely these fields.
The term “agentic AI” matters here.
Put simply, it refers to AI systems that do more than answer a single question. They may pursue an objective through several steps, draw on different sources of information or tools, and decide what action should come next.
When such a system completes an office task, the question is usually one of efficiency.
But when technology begins to interact with a person's voice, gestures, emotions or creative signals, the question changes.
How deeply will machines understand us?
The more difficult question is this:
Will we understand how machines are changing us?
‘Modes of Vibration’: Not an Album, but a System
Sassoon's 2025 work, Modes of Vibration, is an artistic exploration of that question.
The project combines neoclassical music, sound sculpture and specially designed AI agents. According to AI for Good, it premiered as the closing audiovisual performance at the 2025 summit.
But its significance lies not merely in its technical complexity.
Sassoon uses sound as though music were not a fixed composition but a living relationship. Human-generated signals, musical movement, possible emotional cues and machine responses are drawn into interaction.
The artist, in other words, is not simply playing an instrument.
A system is forming a kind of “relationship” with him.
From the performance system behind Modes of Vibration, known as TEAL, Sassoon later developed the Personal Construct, a five-layer agentic operating system. According to his AI for Good profile, he plans to present it publicly at the 2026 summit.
The name “Personal Construct” itself points toward a larger question.
In the age of the personal computer, people used machines.
In the age of agentic AI, will machines increasingly construct personalized digital environments around people?
And if they do, who will be the architect of those environments?
The user?
The AI developer?
The training data?
Or algorithmic decisions, most of which remain invisible to the person being affected?
Is Emotion Now Data?
Perhaps the most compelling idea in Sassoon's research is “emotional sovereignty.”
An AI for Good session description presents the idea as a person's right to retain authority over the course of their own emotional life, even in the presence of emotionally responsive technologies. It also warns that emotionally attuned interfaces, combined with large language models, may use subtle feedback loops to influence behaviour, attachment and emotional norms.
It may sound philosophical.
In reality, it is deeply practical.
We already live inside digital systems that know which videos make us pause, which sentences provoke us, which images we click, how late we remain awake and what kind of notification can draw us back to a platform.
AI can make these systems far more personal.
A machine may not merely know what I like. It may infer signals from my tone of voice, hesitation, disappointment, excitement or emotional vulnerability.
That is where an important distinction emerges.
Understanding a person and influencing a person are not the same thing.
But the history of commercial technology suggests that the boundary between the two does not always remain intact.
This is where Sassoon's idea of emotional sovereignty becomes powerful. His argument is that AI should amplify human intention rather than simply extract value from it. His official AI for Good biography identifies this principle as central to his research and refers to his work in the Harvard Business Review, as well as earlier writing in MIT Technology Review and the MIT Computer Music Journal.
But this is also where the difficult questions must begin.
“Amplifying human intention” is an attractive principle. But how does an AI system determine a person's genuine intention?
Is an individual's immediate desire always the same as that person's long-term interest?
If an AI tells users exactly what they want to hear, is it expanding their freedom, or trapping them inside a comfortable digital mirror?
Emotional sovereignty, therefore, cannot remain only an appealing artistic idea.
At some point, it must be translated into design principles, meaningful consent, limits on data use, independent assessment and human oversight.
Otherwise, “human-centred AI” may become just another reassuring slogan of the technology industry.
A Ballet Where AI Is Not Merely Backstage
In 2023, Sassoon's project Fusion offered another experiment in the relationship between AI, music and the human body.
Created with the beatbox artist Reeps One, the Leipzig International Ballet and the Leipzig Opera, the work has been described in Sassoon's AI for Good biography as the first AI-based ballet.
Its significance was not simply that artificial intelligence had been used in a ballet performance.
Dance is, fundamentally, the language of the body.
Music is a language of time.
AI is a technology of pattern recognition and response.
When all three become part of the same system, the conventional relationship between artist and technology begins to change.
If a machine responds to human expression during a live performance, who is following whom?
Is the human following the machine?
Is the machine following the human?
Or are both creating a third kind of creative situation together?
For art, this is an exciting question.
For governance, it is also an alarmingly relevant one.
Where Does Sassoon Fit Within the United Nations' Four Questions?
To regard Gadi Sassoon's performance simply as a cultural interlude would be to miss the larger significance of the Geneva moment.
The first thematic cluster of the Global Dialogue concerns the social, economic, ethical, cultural, linguistic and technical implications of AI.
Sassoon's work enters directly into the cultural and ethical dimensions of that debate. His question, in effect, is this: when AI becomes part of creative expression, what do we mean by a “human” work?
The second cluster focuses on bridging AI divides, capacity-building, access and digital foundations. The United Nations has linked this discussion to capacity gaps in developing countries, high-performance computing and related skills, open-source software, open data and open AI models.
Here, Sassoon's work also reveals an uncomfortable contrast.
On a stage in Geneva, highly advanced agentic AI can be used to experiment with emotion, music and live performance.
Yet many of the world's languages remain poorly represented in major AI systems. Researchers in numerous countries still lack sufficient computing infrastructure. Countless artists do not enjoy equal access to advanced AI tools.
So whose future is the future of “AI-assisted creativity”?
Milan's?
Geneva's?
Silicon Valley's?
Or also the future of Fulbagicha village in Lalmohan Upazila on Bhola Island, Tajmohal Road, Shyamoli, Kigali, Lima, Nawabgonj in Dhaka's Lalbagh and Kathmandu?
The third cluster concerns safe, secure and trustworthy AI. The fourth focuses on human rights, transparency, accountability and robust human oversight.
This is where the idea of emotional sovereignty begins to move beyond the boundaries of art.
If an AI system can infer my emotional state, must I be told?
Where will that inference be stored?
Who will be allowed to use it?
An insurance company?
An employer?
A political campaign?
An educational platform?
And if AI misinterprets my emotions, to whom do I appeal?
se are not questions about music.
They are questions about rights and power.
Art Does Not Provide the Answer. Sometimes It Asks the Right Question.
At the opening session of the Dialogue, António Guterres argued that in high-stakes decisions, machines may provide information, but humans must make the decisions and remain accountable for them.
The principle is relatively easy to understand in a courtroom, a hospital or a policing system.
But what does it mean in the realm of human emotion?
What does “human oversight” look like when an AI speaks with me every day, claims to understand my loneliness, collaborates in my creative work and continuously learns from my responses?
This may be where the greatest value of Sassoon's work lies.
He does not tell us that AI is good.
He does not tell us that AI is bad.
Instead, his performance makes visible the uncertain territory between human and machine, a place where collaboration may gradually become dependence, personalization may become influence and the imitation of empathy may be mistaken for genuine understanding.
Yet we should also resist giving art more responsibility than it can carry.
A beautiful audiovisual performance does not make AI safe.
An attractive philosophical idea is not a substitute for an accountability framework.
The staging of technology can sometimes make its power appear so beautiful that we forget its political meaning.
That warning matters too.
Still, Sassoon's presence in Geneva carried a significance of its own.
At a time when the world speaks about AI governance through the languages of law, safety, computing, markets and geopolitics, an artist reminded us that this technology will also enter human life through the door of feeling.
How intelligent technology becomes is certainly important.
But perhaps the harder question is another one.
In the presence of increasingly intimate and emotionally responsive technologies, how much authority will human beings retain over their own feelings, intentions and creative freedom?
Gadi Sassoon's music does not answer that question.
Good art perhaps does not always provide answers.
Sometimes its task is to ask a question that continues to resonate in our minds long after the conference hall has fallen silent.
A. H. M. Bazlur Rahman is an
Expert on digital democracy,
information integrity and media
development policy. Moreover, he is Ambassador for Responsible Artificial Intelligence and Antimicrobial
Resistance in the United Nations.
