How AI Art Could Enhance Humanity’s Collective Memory | Refik Anadol | TED

Let me begin by saying

that what you are looking at is not real.

These are artificial corals created by a generative AI algorithm,

trained with more than 100 million coral images.

Across the globe,

sea water is becoming less habitable due to the climate change

and the coral reefs are dying rapidly.

One day we might be only left with the simulations of corals

in a virtual world.

With this project, “Coral Dreams,”

our aim is to use AI

and try to create artificial realities while preserving disappearing nature.

I'm a media artist and director.

My team and I have been using generative AI as a collaborator

for seven years,

although it feels like 70.

We train machine-learning algorithms by harnessing large,

focused and publicly available data sets

and visualizing what I call humanity's collective memories,

such as nature, urban and culture.

Since [the] pandemic,

my focus has been to compile the largest data sets

and artificially preserve nature.

I am optimistic about generative AI

because of its potential for enhancing our memories.

We as artists can utilize this potential not only to represent nature,

but also to remember how it feels to be immersed in it in a digital age.

Generative AI creates possibilities to train algorithms with any image, sound,

text and even scent data.

For example, this is “Floral Dreams,”

an algorithm trained with more than 75 million floral images

of 16,000 species.

By using more than half a million scent molecules,

we were able to create the scent of these dreams.

Now let's please imagine a living archive

that we can walk into.

A universe that is constantly reimagined, [reconstructing] its forms, patterns,

colors and scents.

Our life is becoming increasingly rooted in digital worlds,

and the boundaries between physical and virtual,

technology and nature, are blurring.

Generative AI helps us to create new realities

and also project onto reality through possibility space.

Can we go to that space?

Can we fill it with our feelings, our senses?

Large language models are just the beginning

of a long journey of innovations [which] will bring more possibilities.

Soon, I believe we will be exploring hyper models,

text to image, to sound,

to scent, to life.

And a big challenge using generative AI in art

is how to provide models with original data.

For this project, “Glacier Dreams,”

we decided not to use existing models.

Instead, we decided to collect our own image,

sound, scent and climate data.

By traveling to our first destination, Iceland,

we were able to capture the beginning of our own narratives of glaciers.

I also believe that AI's capacity can be mapped

onto the complex history of human wisdom and consciousness in nature.

Could we use AI to preserve and learn about ancient knowledge in nature?

This was one of the first questions in my mind

when I met with the wonderful leaders

of Yawanawa tribe in Brazil, Acre, Amazonia, in the rainforest.

My mentors and heroes, Chief Nixiwaka and his creative force Putanny,

who oversee their cultural preservation and ecological sanctuary.

I became deeply inspired by their ways of learning

and remembering already existing knowledge in nature.

Together we started a new project,

a respectful co-creation and open-source AI rainforest model.

With this model,

generative AI can even reconstruct extinct flora and fauna

based on the tribe's deep and collective knowledge.

This project will help us, hopefully,

to bring ancient wisdom to our society respectfully.

My hope is that one day

AI becomes a mirror that can reflect collective memories of all humanity.

And I do believe that we can use it to bring people of any age and culture,

inspiration, joy and hope.

Thank you.

(Applause)