< Return to Video

Art in the age of machine intelligence

  • 0:01 - 0:04
    Hi, I'm Refik. I'm a media artist.
  • 0:04 - 0:06
    I use data as a pigment
  • 0:06 - 0:08
    and paint with a thinking brush
  • 0:08 - 0:10
    that is assisted
    by artificial intelligence.
  • 0:11 - 0:14
    Using architectural spaces as canvases,
  • 0:14 - 0:16
    I collaborate with machines
  • 0:16 - 0:19
    to make buildings dream and hallucinate.
  • 0:19 - 0:21
    You may be wondering,
    what does all this mean?
  • 0:21 - 0:25
    So let me please take you
    into my work and my world.
  • 0:26 - 0:30
    I witnessed the power of imagination
    when I was eight years old,
  • 0:30 - 0:32
    as a child growing up in Istanbul.
  • 0:32 - 0:35
    One day, my mom brought home
    a videocassette
  • 0:35 - 0:38
    of the science-fiction movie
    "Blade Runner."
  • 0:38 - 0:41
    I clearly remember being mesmerized
  • 0:41 - 0:46
    by the stunning architectural vision
    of the future of Los Angeles,
  • 0:46 - 0:49
    a place that I had never seen before.
  • 0:49 - 0:54
    That vision became
    a kind of a staple of my daydreams.
  • 0:54 - 0:57
    When I arrived in LA in 2012
  • 0:57 - 0:59
    for a graduate program
    in Design Media Arts,
  • 0:59 - 1:02
    I rented a car and drove downtown
  • 1:02 - 1:05
    to see that wonderful world
    of the near future.
  • 1:05 - 1:07
    I remember a specific line
  • 1:07 - 1:10
    that kept playing
    over and over in my head:
  • 1:10 - 1:12
    the scene when the android Rachael
  • 1:12 - 1:16
    realizes that her memories
    are actually not hers,
  • 1:16 - 1:21
    and when Deckard tells her
    they are someone else's memories.
  • 1:21 - 1:22
    Since that moment,
  • 1:22 - 1:25
    one of my inspirations
    has been this question.
  • 1:26 - 1:30
    What can a machine do
    with someone else's memories?
  • 1:30 - 1:33
    Or, to say that in another way,
  • 1:33 - 1:36
    what does it mean to be an AI
    in the 21st century?
  • 1:37 - 1:40
    Any android or AI machine
  • 1:40 - 1:43
    is only intelligent
    as long as we collaborate with it.
  • 1:43 - 1:45
    It can construct things
  • 1:45 - 1:48
    that human intelligence intends to produce
  • 1:48 - 1:50
    but does not have the capacity to do so.
  • 1:51 - 1:55
    Think about your activities
    and social networks, for example.
  • 1:55 - 1:59
    They get smarter
    the more you interact with them.
  • 1:59 - 2:03
    If machines can learn or process memories,
  • 2:03 - 2:05
    can they also dream?
  • 2:05 - 2:07
    Hallucinate?
  • 2:07 - 2:09
    Involuntarily remember,
  • 2:09 - 2:13
    or make connections
    between multiple people's dreams?
  • 2:13 - 2:19
    Does being an AI in the 21st century
    simply mean not forgetting anything?
  • 2:20 - 2:22
    And, if so,
  • 2:22 - 2:26
    isn't it the most revolutionary thing
    that we have experienced
  • 2:26 - 2:31
    in our centuries-long effort
    to capture history across media?
  • 2:31 - 2:33
    In other words,
  • 2:33 - 2:36
    how far have we come
    since Ridley Scott's "Blade Runner"?
  • 2:37 - 2:40
    So I established my studio in 2014
  • 2:40 - 2:42
    and invited architects,
  • 2:42 - 2:45
    computer and data scientists,
    neuroscientists,
  • 2:45 - 2:47
    musicians and even storytellers
  • 2:47 - 2:50
    to join me in realizing my dreams.
  • 2:51 - 2:54
    Can data become a pigment?
  • 2:54 - 2:56
    This was the very first question we asked
  • 2:56 - 3:01
    when starting our journey
    to embed media arts into architecture,
  • 3:01 - 3:04
    to collide virtual and physical worlds.
  • 3:05 - 3:09
    So we began to imagine
    what I would call the poetics of data.
  • 3:10 - 3:13
    One of our first projects,
    "Virtual Depictions,"
  • 3:13 - 3:15
    was a public data sculpture piece
  • 3:15 - 3:18
    commissioned by the city of San Francisco.
  • 3:18 - 3:20
    The work invites the audience
  • 3:20 - 3:23
    to be part of a spectacular
    aesthetic experience
  • 3:23 - 3:25
    in a living urban space
  • 3:25 - 3:30
    by depicting a fluid network
    of connections of the city itself.
  • 3:30 - 3:33
    It also stands as a reminder
  • 3:33 - 3:37
    of how invisible data
    from our everyday lives,
  • 3:37 - 3:40
    like the Twitter feeds
    that are represented here,
  • 3:40 - 3:42
    can be made visible
  • 3:42 - 3:48
    and transformed into sensory knowledge
    that can be experienced collectively.
  • 3:49 - 3:54
    In fact, data can only become knowledge
    when it's experienced,
  • 3:54 - 3:58
    and what is knowledge and experience
    can take many forms.
  • 3:58 - 3:59
    When exploring such connections
  • 3:59 - 4:04
    through the vast potential
    of machine intelligence,
  • 4:04 - 4:09
    we also pondered the connection
    between human senses
  • 4:09 - 4:13
    and the machines' capacity
    for simulating nature.
  • 4:13 - 4:18
    These inquiries began
    while working on wind-data paintings.
  • 4:18 - 4:21
    They took the shape of visualized poems
  • 4:21 - 4:26
    based on hidden data sets
    that we collected from wind sensors.
  • 4:26 - 4:28
    We then used generative algorithms
  • 4:28 - 4:33
    to transform wind speed,
    gust and direction
  • 4:33 - 4:35
    into an ethereal data pigment.
  • 4:36 - 4:40
    The result was a meditative
    yet speculative experience.
  • 4:41 - 4:44
    This kinetic data sculpture,
    titled "Bosphorus,"
  • 4:44 - 4:49
    was a similar attempt to question
    our capacity to reimagine
  • 4:49 - 4:50
    natural occurrences.
  • 4:51 - 4:56
    Using high-frequency radar collections
    of the Marmara Sea,
  • 4:56 - 4:58
    we collected sea-surface data
  • 4:58 - 5:01
    and projected its dynamic movement
    with machine intelligence.
  • 5:02 - 5:04
    We create a sense of immersion
  • 5:04 - 5:08
    in a calm yet constantly changing
    synthetic sea view.
  • 5:10 - 5:14
    Seeing with the brain
    is often called imagination,
  • 5:14 - 5:16
    and, for me, imagining architecture
  • 5:16 - 5:20
    goes beyond just glass, metal or concrete,
  • 5:20 - 5:24
    instead experimenting with
    the furthermost possibilities of immersion
  • 5:24 - 5:29
    and ways of augmenting
    our perception in built environments.
  • 5:29 - 5:32
    Research in artificial intelligence
    is growing every day,
  • 5:32 - 5:36
    leaving us with the feeling
    of being plugged into a system
  • 5:36 - 5:38
    that is bigger and more knowledgeable
  • 5:38 - 5:40
    than ourselves.
  • 5:40 - 5:44
    In 2017, we discovered
    an open-source library
  • 5:44 - 5:46
    of cultural documents in Istanbul
  • 5:46 - 5:50
    and began working on "Archive Dreaming,"
  • 5:50 - 5:54
    one of the first AI-driven
    public installations in the world,
  • 5:54 - 6:01
    an AI exploring approximately
    1.7 million documents that span 270 years.
  • 6:02 - 6:05
    One of our inspirations
    during this process
  • 6:05 - 6:08
    was a short story
    called "The Library of Babel"
  • 6:08 - 6:11
    by the Argentine writer Jorge Luis Borges.
  • 6:11 - 6:17
    In the story, the author conceives
    a universe in the form of a vast library
  • 6:17 - 6:23
    containing all possible 410-page books
    of a certain format and character set.
  • 6:23 - 6:24
    Through this inspiring image,
  • 6:25 - 6:29
    we imagine a way to physically explore
    the vast archives of knowledge
  • 6:29 - 6:32
    in the age of machine intelligence.
  • 6:32 - 6:34
    The resulting work, as you can see,
  • 6:34 - 6:36
    was a user-driven immersive space.
  • 6:37 - 6:41
    "Archive Dreaming" profoundly transformed
    the experience of a library
  • 6:41 - 6:44
    in the age of machine intelligence.
  • 6:44 - 6:48
    "Machine Hallucination"
    is an exploration of time and space
  • 6:48 - 6:53
    experienced through New York City's
    public photographic archives.
  • 6:53 - 6:55
    For this one-of-a-kind immersive project,
  • 6:55 - 6:58
    we deployed machine-learning algorithms
  • 6:58 - 7:02
    to find and process over
    100 million photographs of the city.
  • 7:03 - 7:06
    We designed an innovative narrative system
  • 7:06 - 7:12
    to use artificial intelligence
    to predict or to hallucinate new images,
  • 7:12 - 7:16
    allowing the viewer
    to step into a dreamlike fusion
  • 7:16 - 7:18
    of past and future New York.
  • 7:20 - 7:22
    As our projects delve deeper
  • 7:22 - 7:25
    into remembering
    and transmitting knowledge,
  • 7:25 - 7:30
    we thought more about how memories
    were not static recollections
  • 7:30 - 7:34
    but ever-changing
    interpretations of past events.
  • 7:34 - 7:36
    We pondered how machines
  • 7:36 - 7:40
    could simulate unconscious
    and subconscious events,
  • 7:40 - 7:44
    such as dreaming,
    remembering and hallucinating.
  • 7:45 - 7:48
    Thus, we created "Melting Memories"
  • 7:48 - 7:51
    to visualize the moment of remembering.
  • 7:52 - 7:54
    The inspiration came from a tragic event,
  • 7:55 - 7:58
    when I found out that my uncle
    was diagnosed with Alzheimer's.
  • 8:00 - 8:02
    At that time, all I could think about
  • 8:02 - 8:07
    was to find a way to celebrate
    how and what we remember
  • 8:07 - 8:09
    when we are still able to do so.
  • 8:09 - 8:13
    I began to think of memories
    not as disappearing
  • 8:13 - 8:16
    but as melting or changing shape.
  • 8:16 - 8:18
    With the help of machine intelligence,
  • 8:18 - 8:22
    we worked with the scientists
    at the Neuroscape Laboratory
  • 8:22 - 8:24
    at the University of California,
  • 8:24 - 8:29
    who showed us how to understand
    brain signals as memories are made.
  • 8:29 - 8:34
    Although my own uncle was losing
    the ability to process memories,
  • 8:34 - 8:38
    the artwork generated by EEG data
  • 8:38 - 8:41
    explored the materiality of remembering
  • 8:41 - 8:46
    and stood as a tribute
    to what my uncle had lost.
  • 8:49 - 8:51
    Almost nothing about contemporary LA
  • 8:52 - 8:55
    matched my childhood
    expectation of the city,
  • 8:55 - 8:58
    with the exception
    of one amazing building:
  • 8:58 - 9:02
    the Walt Disney Concert Hall,
    designed by Frank Gehry,
  • 9:02 - 9:04
    one of my all-time heroes.
  • 9:04 - 9:08
    In 2018, I had a call
    from the LA Philharmonic
  • 9:08 - 9:10
    who was looking for an installation
  • 9:10 - 9:14
    to help mark the celebrated symphony's
    hundred-year anniversary.
  • 9:14 - 9:17
    For this, we decided to ask the question,
  • 9:17 - 9:20
    "Can a building learn? Can it dream?"
  • 9:21 - 9:22
    To answer this question,
  • 9:22 - 9:28
    we decided to collect everything recorded
    in the archives of the LA Phil and WDCH.
  • 9:28 - 9:33
    To be precise, 77 terabytes
    of digitally archived memories.
  • 9:33 - 9:35
    By using machine intelligence,
  • 9:35 - 9:38
    the entire archive, going back 100 years,
  • 9:38 - 9:41
    became projections on the building's skin,
  • 9:41 - 9:45
    42 projectors to achieve
    this futuristic public experience
  • 9:45 - 9:47
    in the heart of Los Angeles,
  • 9:47 - 9:51
    getting one step closer
    to the LA of "Blade Runner."
  • 9:52 - 9:55
    If ever a building could dream,
  • 9:55 - 9:56
    it was in this moment.
  • 10:00 - 10:04
    Now, I am inviting you to one last journey
    into the mind of a machine.
  • 10:06 - 10:09
    Right now, we are fully immersed
    in the data universe
  • 10:09 - 10:14
    of every single curated TED Talk
    from the past 30 years.
  • 10:14 - 10:20
    That means this data set includes
    7,705 talks from the TED stage.
  • 10:21 - 10:26
    Those talks have been translated
    into 7.4 million seconds,
  • 10:26 - 10:30
    and each second is represented
    here in this data universe.
  • 10:30 - 10:32
    Every image that you are seeing in here
  • 10:32 - 10:35
    represents unique moments
    from those talks.
  • 10:35 - 10:37
    By using machine intelligence,
  • 10:37 - 10:41
    we processed a total of 487,000 sentences
  • 10:41 - 10:46
    into 330 unique clusters of topics
    like nature, global emissions,
  • 10:46 - 10:49
    extinction, race issues, computation,
  • 10:49 - 10:52
    trust, emotions, water and refugees.
  • 10:52 - 10:55
    These clusters are then
    connected to each other
  • 10:55 - 10:57
    by an algorithm,
  • 10:57 - 11:00
    [that] generated 113 million
    line segments,
  • 11:01 - 11:04
    which reveal new conceptual relationships.
  • 11:04 - 11:07
    Wouldn't it be amazing
    to be able to remember
  • 11:07 - 11:10
    all the questions that have ever
    been asked on the stage?
  • 11:12 - 11:13
    Here I am,
  • 11:13 - 11:16
    inside the mind
    of countless great thinkers,
  • 11:16 - 11:20
    as well as a machine,
    interacting with various feelings
  • 11:20 - 11:22
    attributed to learning,
  • 11:22 - 11:24
    remembering, questioning
  • 11:24 - 11:28
    and imagining all at the same time,
  • 11:28 - 11:30
    expanding the power of the mind.
  • 11:31 - 11:33
    For me, being right here
  • 11:33 - 11:37
    is indeed what it means
    to be an AI in the 21st century.
  • 11:38 - 11:40
    It is in our hands, humans,
  • 11:41 - 11:44
    to train this mind to learn and remember
  • 11:44 - 11:46
    what we can only dream of.
  • 11:47 - 11:48
    Thank you.
Title:
Art in the age of machine intelligence
Speaker:
Refik Anadol
Description:

What does it look like inside the mind of a machine? Inspired by the architectural vision of a futuristic Los Angeles in "Blade Runner," media artist Refik Anadol melds art with artificial intelligence in his studio's collaborations with architects, data scientists, neuroscientists, musicians and more. Witness otherworldly installations that might make you rethink the future of tech and creativity.

more » « less
Video Language:
English
Team:
closed TED
Project:
TEDTalks
Duration:
12:01

English subtitles

Revisions Compare revisions