00:00Bonjour, bienvenue, je m'appelle Romain Simeon, je suis ici sur le nom de l'IRCAM Amplify company.
00:11Nous sommes une société qui croient fortement en la power de son, que vous êtes actuellement experiencing.
00:20Nous croire que son son enhance et créent de nouvelles expériences pour tous et pour toutes les vies.
00:33Aujourd'hui, je vais parler de l'IRCAM entre l'IRCAM et la musique entre l'IRCAM et l'IRCAM.
00:42L'IRCAM est un company de l'IRCAM, et nous sommes attachés à l'IRCAM de l'IRCAM en Paris,
00:50France.
00:51Je voudrais faire un peu d'experiment.
00:54Pour un moment, could you please close your eyes and listen for just a minute.
01:01In the dawn of a digital age
01:07We're melodies composed on a binary page
01:12A symphony woven by circuits and wires
01:17The birth of music
01:20The soul of the silent pyre
01:25Lines of code intricate and grand
01:28Birthing harmonies with an artificial hand
01:33The eye conductor
01:36Guiding the soul
01:38Creating emotions
01:40Right where it belongs
01:44From ones and zeros
01:48Now
01:52Melodies evolve
01:54The heartbeat of technology
01:57We all resolve
01:59Yet
02:00In this symphony
02:04Something feels amiss
02:08A soulless
02:10You can open your eyes now
02:12Can anyone tell me who was the artist?
02:15Raise your hands
02:17Anyone recognize?
02:20Alright, that's a good answer
02:22Because actually the artist is Suno
02:25An AI-generative engine for music
02:30And the prompt I use to generate that song
02:32Is a mellow folk song
02:34About the progress of AI music
02:38Did you like it?
02:41Yeah?
02:42I see, yeah
02:42Cool
02:46Okay
02:47Now I'm going to talk
02:48I'm going to give you a little bit of fun facts
02:50About AI and the music industry
02:52To generate this song
02:55It took me less than one minute
02:57To type the prompt
02:58Wait for the generative engine
03:00To give me the file
03:02And download it
03:04In 2023
03:06Over 170 million tracks
03:11Were generated by AI in the world
03:16There's a playlist on Spotify
03:18Which I will not give you the name
03:20Because I don't want to
03:21That has around 240,000 listeners monthly
03:27And the top track on this playlist
03:31I think it's about 500,000 plays
03:38Or listeners basically
03:40Which is pretty crazy and impressive, right?
03:45Technology goes super fast
03:48So there is a very obvious question
03:52And something that we try
03:53At IRCAM Amplify to answer is
03:55Should we be afraid of what's going on
03:58In the music industry
03:59And in the music world
04:01And in the artistic world
04:02Regarding AI?
04:05First of all
04:06Let's have a quick look
04:07At where AI today stands
04:09Regarding music
04:11So as you can see
04:12AI is everywhere
04:15From professional tools
04:17To very end consumer tools
04:19With also a huge amount
04:21Of foundations models
04:22This is great
04:23This fosters innovation
04:24This goes really fast
04:26It's fantastic
04:30Why is there a special bond?
04:32And why this question
04:34Regarding AI in music today?
04:36Because art has always been
04:41A huge vector for innovation
04:44And has always been extremely keen
04:46To use AI
04:48And use technical tools
04:50To innovate and create new things
04:53So when it comes to creation
04:56To emotion
04:57And to ideation
04:58To inventing new things
05:00AI seems like the perfect tool for music
05:08There's one thing
05:10That's important to notice
05:11Is that AI is a very logical tool
05:14It's very first in, first out
05:16Whereas music and art in general
05:19Is all about emotion
05:20So there's always this need for guidance
05:24Between these two fields
05:26Most specifically
05:29There are a huge amount of challenges
05:32When it comes to the current state
05:34Of gen AI and music
05:35So let's talk about this
05:37For a little bit
05:40Three topics
05:41Legal and ethical
05:43Technical and artistic
05:46From a legal perspective
05:47You might have heard of this recently
05:50Sony has sent
05:52I think to six or seven hundred companies
05:55Yeah, seven hundred
05:58A warning
05:59Stop using our content
06:01Without our authorization
06:03This is a huge challenge today
06:06For gen AI
06:07In the music industry
06:08Why are you using our content?
06:14Second, technical
06:15This one is a little bit easier
06:17To understand
06:18And to basically bypass
06:19Time is going to go its way
06:23The cost to generate
06:25To operate
06:28The size of the data sets
06:30That are used to train these models
06:32Needs to get smaller and smaller
06:36But there's also a technical challenge
06:38When it comes to quality
06:39I don't know if you've heard it
06:41Because honestly
06:41It's quite loud here
06:43But the Suno sound that you heard
06:47Is of pretty poor quality
06:49And I'm not judging the content
06:51I'm really judging the file itself
06:52So if you're using title
06:54That's not going to work
06:57And finally
06:58And this one for all of us
07:00At Surcam Amplify
07:00Is the most important
07:01The artistic part of it
07:03Today when you
07:05Use a prompt
07:06Gen AI
07:08Basically you type in a sentence
07:10And you get a result
07:11There is no artistic guidance
07:14No emotion
07:16We were just talking
07:17Like a few minutes earlier
07:19With our CTO
07:20About the fact that
07:21Basically when you
07:23Want to generate something
07:26When you want to write something
07:27When you want to create something
07:28There is a part of you
07:30A part of your emotion
07:31A part of your feelings
07:32That goes into that
07:33When you ask for an AI
07:35It's going to be
07:36Really the basic stuff
07:37And this is where
07:39For us at Amplify
07:40We think
07:42The emotion
07:43Is a huge requirement
07:44When it comes to Gen AI
07:46And this is where
07:46The biggest gap is today
07:49Go test Suno
07:50And try to
07:51You know
07:52Put something from you
07:54In there
07:55Impossible
07:59So
08:00Being afraid of AI
08:01Is okay
08:03And
08:03Yeah
08:04I mean
08:05Some think
08:06That at some point
08:07Gen AI
08:08Is going to replace artists
08:09Why do we think
08:10Why do I think
08:11That's not going to happen
08:12I've kind of gave you
08:14Some hints
08:15Previously
08:15But
08:16Basically because
08:17We think that
08:18The best position
08:19For Gen AI
08:20In the future
08:21When it comes to music
08:22Is to become
08:23A real tool
08:24A tech enabler
08:26To become
08:28A co-pilot
08:30And
08:31Yeah
08:31A co-producer
08:32To help
08:33The artist
08:34Enhance the artist
08:36We
08:37Are also sure
08:39That in terms of
08:40Creativity
08:41In terms of
08:42Visibility
08:43In terms of
08:44Discoverability
08:45Gen AI
08:45Is going to be
08:46A fantastic tool
08:47And this is what
08:48We are
08:49Working on today
08:50We are actually
08:52You know
08:52Doing a lot of stuff
08:54At Yerkam Amplified
08:54I'll tell you about
08:55In a second
08:55And finally
08:57In terms of immersion
08:58And in terms of emotion
09:00Gen AI
09:01Can become
09:02Something fantastic
09:03When it comes to
09:04For example
09:04Look at Spotify's
09:05AI playlists
09:06This is awesome
09:07This works really well
09:08And we are working
09:10On the exact same
09:11Engine
09:11Available to everyone
09:13To actually
09:14Have a Gen AI
09:15That generates
09:16Playlists
09:17With all the tools
09:18That we at Yerkam Amplify
09:19Provide to the world
09:24We live in a
09:25Very complex world
09:27That is
09:28Extremely
09:29Sensitive
09:30And
09:31I think it's
09:32The right time
09:33To quote
09:33Frank Herbert
09:35From June
09:35This is
09:36The exact
09:36First sentence
09:37Of the first book
09:39A beginning
09:40Is the time
09:41For taking
09:41The most delicate
09:42Care
09:43That the balances
09:44Are correct
09:44And I think
09:45This
09:45Today
09:46Is what
09:47We all need
09:48To do
09:49Looking at
09:50Gen AI
09:51As a
09:52Cool tool
09:53Or a cool
09:54Technology
09:54Is fantastic
09:55Looking at it
09:56Today
09:57As a
09:58Creator
09:58Is
10:00In our opinion
10:00Not the right
10:01Way
10:02To look
10:02At it
10:06We are
10:07Earkam Amplify
10:08We provide
10:10All kind
10:11Of tools
10:12On the
10:12Dot
10:13I.O.
10:13Platform
10:14That you're
10:15Seeing here
10:15We think
10:16It's time
10:17To listen
10:18For all of us
10:19And to the world
10:20Thank you
10:21Very much
Commentaires