06/04/2026 | News release | Archived content
At the opening of the CISAC General Assembly 2026 in Paris, CISAC President and ABBA co-founder Björn Ulvaeus, delivered a speech on the future of human creativity in the age of artificial intelligence.
Watch the full recording of the speech here.
The full text of Björn Ulvaeus' address is reproduced below.
______
Good morning everyone.
It's a big day and we are living in momentous times.
I'd like to start this speech with a big question. A philosophical question that I think will prove fundamental for the organisations that we represent. And one that if we are honest, none of us can answer with complete confidence.
Does the source of art matter?
If a piece of music moves you - genuinely moves you, reaches something real inside you - does it matter whether a human made it?
If you close your eyes and something passes through you - some recognition of grief, of joy, or longing, and you later discover it was assembled by a machine - does that change what happened to you? Does it unhappen?
I used to think the answer was obviously yes.
Now, I'm not so sure. And I think we need to be honest about that uncertainty rather than arguing past it as though it doesn't exist.
I'll be honest from the start: I use AI in my songwriting and I think it's a fantastic tool. It opens up new possibilities, helps me explore ideas faster and can be a genuinely creative collaborator.
So, I'm not here to reject technology. I'm here because we must try to understand what we are actually dealing with.
At Davos in January, Yuval Noah Harari, the historian, said something that has stayed with me. He said that AI will beat him. It will beat him at the thing he has built his entire life around: putting words in order.
He's an author, a speaker - arranging words is his game. And AI will beat him at it. He doesn't know if it takes two years or ten, but it will.
His argument has to do with structure. If thinking is largely language-based, then AI already outperforms many humans. Which means anything made of words will increasingly be taken over by AI.
His conclusion is stark.
Whether humans retain a place in that world depends on the value we assign to our nonverbal feelings and our embodied life experiences - wisdom, if you like, that cannot be fully expressed in words.
If we continue to define ourselves by our ability to think in language, our identity will begin to erode.
That's unsettling because language is what we do.
Songs are language, stories are language. The art forms most of us in this room have spent our lives making and protecting are built from words and the spaces between them.
Then again, songs are not only words, they are music as well.
So, what is music exactly? Is it a language? And if it is - or if it isn't - what follows from that?
This turns out to be one of the oldest and most contested questions in philosophy and neuroscience. And the answer is: music is both and neither, and the ambiguity is the point in there.
Music has structure. It has patterns and rules that a trained ear recognises, much like grammar in language. In that sense, there are clear parallels between music and language, and this is precisely the part that AI already mastered.
But music is not only structure.
Steven Pinker argues that while music shares the structural features of language, it lacks its semantics. It cannot communicate specific meaning on its own. A sentence tells you something. A melody makes you feel something, but it does not tell you what to feel it about.
So, Harari says the last human refuge, the territory AI cannot fully colonize, lies in non-verbal feeling and embodied experience. And music, it turns out, lives precisely there. Not entirely in language, not entirely outside it, it sits on that fault line.
A human domain, at last. Or so we would like to think.
But I have to be honest with you about something uncomfortable.
The machine is already reaching into that space. It already knows how to manipulate our feelings.
In a recent study, AI-generated music triggered stronger emotional responses than human-composed music. Participants described the AI music as more exciting, even if human music felt more familiar.
More exciting.
In a blind test, the machine wins in the very field - emotional impact - that we assumed was ours alone.
So, if human creativity is not defined by superior quality, or greater emotional power, or even originality, because machine may match us there, what is left?
I think the answer is this:
Human creativity is not just expression. It is testimony. Testimony of a life lived.
A human being who writes a song about grief has grieved. One who writes about love has loved, and very probably lost. The song is not just a product. It is evidence. Evidence that something happened to a living person, that they enjoyed or endured, and found a way to make transmissible to others.
That is what art has always been. Not decoration. Testimony.
Harari puts this in terms that feel almost unbearably precise. He invokes the ancient tension between word and flesh, between that which can be expressed in language and that which lies beyond words.
AI can read every poem ever written and describe the feeling of love more eloquently than any poet. But these are still just words.
It's the map, not the territory. The symbol, not the thing.
We can only hope that humans will continue to care about the territory, even if the map becomes extraordinarily beautiful.
And if we believe that, as I do, then this is no longer only a philosophical question. It becomes a political one and a legal one.
Because beliefs like these do not enforce themselves. They require structure. They require law.
Testimony has always needed a witness, and witnesses need protection. On that front, the news is mixed.
In March, the UK government scrapped its proposal to require creators to actively opt out of having their work mined by AI following strong resistance from across the creative sector.
That was a real victory won by people who understood what was at stake and refused to be silent.
On July 31, the Munich court will rule in GEMA versus Suno - Europe's first major test of whether training generative AI on copyright music is infringement.
And in a Massachusetts courtroom this summer, a judge will rule on whether Suno's use was fair. If Suno wins, every licensing deal in the
AI music space collapses. If it loses, licensing becomes the law of the land.
A historic fork in the road, and we are standing at it now.
So, let me return to where I began. Does the source of art matter?
I believe it does.
Not because humans can arrange sounds more cleverly, but because we arrange them having lived, having been afraid, having loved, having lost people, having stood in front of something we could not explain, and tried anyway to find words, or notes, or the silence in the right place.
Music, perhaps of all art forms, lives closest to that experience. It reaches back before language, into the part of us that was moved before we had words to explain why.
That is its miracle. And that's the thing the machine does not have.
It has no skin in the game, no stake in the answer to my original question. It does not lie awake wondering whether it matters.
It will not mourn if the answer turns out to be no.
If we can no longer detect the value of the human hand behind the work. If the source stops mattering…
I would mourn. We all would.
And that, I think, is where the argument begins.
People like us in this room have spent a century insisting - in legal and practical terms - that the person behind the work is real, identifiable and owed something.
That insistence has never been more important. And it has never been more tested.
What we do in the next few years will determine whether when this settles, there is still an ecosystem in which human creators can exist, make a living, and pass something of themselves to those who come after. Whether they do it in collaborating relation with AI or not.
That is the work.
It has always been the work.
Happy 100th birthday, CISAC.
Thank you.