1
00:00:00,600 --> 00:00:02,800
Speaker 1: I think a lot of people think it's just going

2
00:00:02,880 --> 00:00:04,760
Speaker 1: to be great. If we just make these things, they're

3
00:00:04,800 --> 00:00:06,320
Speaker 1: just going to be great, and we're fine. And a

4
00:00:06,320 --> 00:00:08,840
Speaker 1: lot of people think this is definitely a dystopia, and

5
00:00:08,840 --> 00:00:12,000
Speaker 1: we're definitely for it, and I think it can truly

6
00:00:12,119 --> 00:00:13,280
Speaker 1: be a third thing.

7
00:00:14,680 --> 00:00:18,400
Speaker 2: Welcome to Crash Course, a podcast about business, political, and

8
00:00:18,440 --> 00:00:21,680
Speaker 2: social disruption and what we can learn from it. I'm

9
00:00:21,720 --> 00:00:28,040
Speaker 2: Tim O'Brien. Today's Crash Course Artificial Intelligence versus the Music Industry.

10
00:00:29,160 --> 00:00:31,200
Speaker 2: The voice you heard at the open was producer and

11
00:00:31,280 --> 00:00:36,080
Speaker 2: songwriter Extraordinary Grimes speaking to us about the good, the bad,

12
00:00:36,280 --> 00:00:40,280
Speaker 2: and the in between of artificial intelligence and music. The

13
00:00:40,360 --> 00:00:43,159
Speaker 2: AI hype may be starting to die down overall, but

14
00:00:43,280 --> 00:00:47,120
Speaker 2: AI isn't going away. You can literally hear it in music.

15
00:00:47,760 --> 00:00:51,360
Speaker 2: Back in April, a super convincing AI duet between rapper

16
00:00:51,440 --> 00:00:55,080
Speaker 2: Drake and singer The Weekend went viral, causing panic in

17
00:00:55,120 --> 00:00:58,440
Speaker 2: the music industry, and it opened the floodgates for deep

18
00:00:58,480 --> 00:01:01,800
Speaker 2: fakes like Frank Sinatra singh hip hop and a fictional

19
00:01:01,840 --> 00:01:07,360
Speaker 2: Oasis mixtape called Aisis. It's not just novelty memes either.

20
00:01:08,000 --> 00:01:11,839
Speaker 2: AI generated background music is sneaking into your Spotify playlist,

21
00:01:12,319 --> 00:01:15,920
Speaker 2: cracking out chill soundscapes, and helping the Beatles put out

22
00:01:15,959 --> 00:01:19,960
Speaker 2: one last song called Now and Then. So music is

23
00:01:20,000 --> 00:01:23,160
Speaker 2: again on the cusp of huge tech driven change. Twenty

24
00:01:23,240 --> 00:01:27,600
Speaker 2: years after file sharing misfit Nabster made CD collections obsolete.

25
00:01:28,440 --> 00:01:30,640
Speaker 2: To help us delve into what that change might look like,

26
00:01:31,200 --> 00:01:34,440
Speaker 2: we at Bloomberg Opinion set one of our columnist detectives,

27
00:01:34,760 --> 00:01:38,920
Speaker 2: Lionelle Laurent, out to investigate. Lionelle is a writer who

28
00:01:38,920 --> 00:01:41,640
Speaker 2: wears many hats, one of which involves keeping an eye

29
00:01:41,680 --> 00:01:45,880
Speaker 2: on digital innovation and upheaval in the music industry, and

30
00:01:45,920 --> 00:01:50,400
Speaker 2: he has a tale to tell. Hi, Lianel, Hi, sim

31
00:01:50,440 --> 00:01:52,360
Speaker 2: I'll let our listeners know that you've been working on

32
00:01:52,400 --> 00:01:56,440
Speaker 2: this from Paris, So before we get started, why music well.

33
00:01:56,640 --> 00:01:59,360
Speaker 3: Music is a personal passion of mine, and music is

34
00:01:59,400 --> 00:02:02,120
Speaker 3: part of what makes as humans, So seeing AI take

35
00:02:02,120 --> 00:02:05,000
Speaker 3: on more of that is I think a big cultural moment.

36
00:02:05,520 --> 00:02:07,920
Speaker 3: But music as an industry has also served as a

37
00:02:08,000 --> 00:02:11,160
Speaker 3: kind of canarian, the coal mine for wider tech disruption,

38
00:02:11,440 --> 00:02:14,080
Speaker 3: including streaming. It really holds a mirror up to the

39
00:02:14,080 --> 00:02:16,240
Speaker 3: rest of the economy, so I think it's important for

40
00:02:16,280 --> 00:02:18,120
Speaker 3: everyone to follow this well.

41
00:02:18,240 --> 00:02:19,840
Speaker 2: I'm going to hand it over to you to tell

42
00:02:19,880 --> 00:02:21,800
Speaker 2: us what you've learned, and I'll catch up with you

43
00:02:21,880 --> 00:02:22,560
Speaker 2: later in the show.

44
00:02:23,080 --> 00:02:26,480
Speaker 3: Thanks Tim. Now, I'm willing to bet that you probably

45
00:02:26,480 --> 00:02:30,040
Speaker 3: haven't heard of French songwriter bunoits Care. He's got a

46
00:02:30,080 --> 00:02:33,840
Speaker 3: special claim to music fame. He's behind the first ever

47
00:02:33,919 --> 00:02:38,200
Speaker 3: song composed by artificial intelligence. It's called Daddy's Car and

48
00:02:38,280 --> 00:02:40,040
Speaker 3: it was released in twenty sixteen.

49
00:02:58,520 --> 00:03:04,200
Speaker 2: In Daddy's Car Sound. So do we like something new?

50
00:03:04,760 --> 00:03:05,400
Speaker 1: It ms?

51
00:03:07,320 --> 00:03:10,120
Speaker 3: This isn't one hundred percent AI. Bern Noir wrote the

52
00:03:10,160 --> 00:03:13,040
Speaker 3: lyrics and arranged the song, and that's him singing, but

53
00:03:13,080 --> 00:03:16,160
Speaker 3: the rest was composed by an AI model trained entirely

54
00:03:16,240 --> 00:03:19,960
Speaker 3: on songs by the Beatles, using all those Fab four classics.

55
00:03:20,240 --> 00:03:22,840
Speaker 3: The machine came up with its own suggestions and ideas.

56
00:03:23,320 --> 00:03:25,560
Speaker 3: It's like if John, Paul, George and Ringo had a

57
00:03:25,560 --> 00:03:29,560
Speaker 3: fifth robotic friend jamming away and Bernoit playing with the results.

58
00:03:31,960 --> 00:03:34,600
Speaker 4: I played with the tool. I fed it with fifty

59
00:03:34,760 --> 00:03:37,760
Speaker 4: songs of the Beatles, my favorite songs of the Beatles,

60
00:03:38,040 --> 00:03:40,880
Speaker 4: and then I got a result eight bars and then

61
00:03:40,880 --> 00:03:43,840
Speaker 4: it was funny, and I continued.

62
00:03:48,960 --> 00:03:51,480
Speaker 3: So I think we can agree this AI music isn't

63
00:03:51,560 --> 00:03:54,280
Speaker 3: quite at the level of let it be Still. Daddy's

64
00:03:54,320 --> 00:03:57,080
Speaker 3: car was a kind of milestone, and for bu Noir

65
00:03:57,240 --> 00:04:00,600
Speaker 3: it is like improvising or writing a regular song hunched

66
00:04:00,640 --> 00:04:03,320
Speaker 3: over a piano. He's been using tech to make music

67
00:04:03,320 --> 00:04:06,000
Speaker 3: since the days of the Commodore sixty four kids ask

68
00:04:06,040 --> 00:04:08,440
Speaker 3: your parents what that is today. He does it all

69
00:04:08,440 --> 00:04:11,960
Speaker 3: from his studio in Paris, a music geeks paradise, stuffed

70
00:04:11,960 --> 00:04:16,040
Speaker 3: with guitars, real real tape and computers. It all helps

71
00:04:16,080 --> 00:04:19,360
Speaker 3: inspire new ideas, which he puts out under the name Skiggit.

72
00:04:19,600 --> 00:04:22,880
Speaker 3: That's Danish for shadow, inspired by a Haunds Christian Anderson

73
00:04:22,920 --> 00:04:33,000
Speaker 3: tale about a shadow that's alive.

74
00:04:36,320 --> 00:04:42,760
Speaker 4: AI helped me to explore an expected part of my creativity.

75
00:04:43,640 --> 00:04:48,400
Speaker 4: It's like the shadow in the Tale of Anderssan with

76
00:04:49,080 --> 00:04:53,560
Speaker 4: becoming something independent from the main character.

77
00:05:04,760 --> 00:05:07,800
Speaker 3: This summer, Bnoir gave me a magical mystery tour of

78
00:05:07,880 --> 00:05:11,320
Speaker 3: how AI could revolutionize music and why it shouldn't be

79
00:05:11,360 --> 00:05:15,440
Speaker 3: banned or dismissed out of hand. First stop deep fakes

80
00:05:15,920 --> 00:05:19,359
Speaker 3: like Heart on My Sleeve that Drake Weekend duet we

81
00:05:19,440 --> 00:05:27,480
Speaker 3: mentioned earlier. Bernoir doesn't think the imitation game is all

82
00:05:27,480 --> 00:05:31,080
Speaker 3: that interesting musically or fair ethically, but he says it

83
00:05:31,120 --> 00:05:33,880
Speaker 3: shows how AI has improved from the days of Daddy's

84
00:05:33,880 --> 00:05:37,560
Speaker 3: Car to do some things incredibly well, like fooling the

85
00:05:37,600 --> 00:05:42,440
Speaker 3: world into thinking a tambra transfer or vocal transfer is genuine.

86
00:05:42,720 --> 00:05:45,680
Speaker 3: Here's Bernoir whipping up a similar kind of song in

87
00:05:45,760 --> 00:05:47,279
Speaker 3: the style of Eminem.

88
00:05:47,640 --> 00:05:50,599
Speaker 4: I asked Judge Appetty to write a song in the

89
00:05:50,680 --> 00:05:54,880
Speaker 4: style of Eminem. I took this this result from jud

90
00:05:54,920 --> 00:06:00,279
Speaker 4: Jipauty and I sung it and it created this result

91
00:06:00,440 --> 00:06:03,440
Speaker 4: with Eminem's timber and you will hear that there is

92
00:06:03,480 --> 00:06:07,120
Speaker 4: a French excellent. The weight of the world is crushing me.

93
00:06:07,160 --> 00:06:09,719
Speaker 5: Down, trying to keep my head up, but I'm feeling

94
00:06:09,760 --> 00:06:13,799
Speaker 5: as well, the thousand fears they're always lurking in my phrasal.

95
00:06:13,960 --> 00:06:15,720
Speaker 6: Song is kind of hurting.

96
00:06:15,920 --> 00:06:19,599
Speaker 3: Wow, that is pretty convincing, though I should mention we're

97
00:06:19,600 --> 00:06:21,280
Speaker 3: not yet at the level of being able to hit

98
00:06:21,279 --> 00:06:24,400
Speaker 3: a button and get instant deep fake back. This took

99
00:06:24,440 --> 00:06:28,080
Speaker 3: several steps. A voice clone, an AI backing track, lyrics

100
00:06:28,080 --> 00:06:31,640
Speaker 3: by chat GPT and Benoir's vocal cords like the so

101
00:06:31,720 --> 00:06:34,680
Speaker 3: called Mechanical Turk, a fake chess playing robot that was

102
00:06:34,720 --> 00:06:36,839
Speaker 3: all the rage and the seventeen hundreds. There is a

103
00:06:36,920 --> 00:06:37,800
Speaker 3: human hidden in there.

104
00:06:38,400 --> 00:06:41,600
Speaker 4: I got in two hours something that is from nothing

105
00:06:41,680 --> 00:06:43,040
Speaker 4: with the voice of Eminem.

106
00:06:43,080 --> 00:06:45,960
Speaker 3: Deep fakes are just the tip of the AI iceberg, though.

107
00:06:46,240 --> 00:06:49,479
Speaker 3: Bunoir also showed us hybridization, a way to mix his

108
00:06:49,600 --> 00:06:52,040
Speaker 3: voice with that of another singer's, in this case the

109
00:06:52,120 --> 00:06:55,000
Speaker 3: late Chet Baker, and create the kind of weird new

110
00:06:55,080 --> 00:06:56,320
Speaker 3: voice in between.

111
00:06:58,000 --> 00:07:01,160
Speaker 4: Seevu chutt.

112
00:07:02,800 --> 00:07:03,560
Speaker 1: That plan.

113
00:07:05,040 --> 00:07:09,840
Speaker 4: Chutts too.

114
00:07:14,440 --> 00:07:17,040
Speaker 3: The next step will be performers able to use these

115
00:07:17,160 --> 00:07:20,760
Speaker 3: kinds of augmented voices on stage, manipulating not just the

116
00:07:20,800 --> 00:07:23,360
Speaker 3: tone of a sinner's voice, but it's feel, it's groove,

117
00:07:23,720 --> 00:07:26,280
Speaker 3: and eventually also have a machine able to do more,

118
00:07:26,400 --> 00:07:29,760
Speaker 3: maybe all of the composing process. I asked Ben, why

119
00:07:29,880 --> 00:07:34,920
Speaker 3: what it's like as a human dealing with the machine.

120
00:07:36,720 --> 00:07:39,760
Speaker 4: The risk is to lose yourself and get bored by

121
00:07:39,840 --> 00:07:43,400
Speaker 4: all the reasons that you get. Because AI is never tied,

122
00:07:44,600 --> 00:07:48,520
Speaker 4: you can always generate, and I must say that often

123
00:07:48,600 --> 00:07:51,280
Speaker 4: lead the results are not so good. Most of the

124
00:07:51,320 --> 00:07:54,160
Speaker 4: results are like average. It takes time to get the

125
00:07:54,200 --> 00:07:57,640
Speaker 4: result that you were expected, or the result that will

126
00:07:57,680 --> 00:07:59,880
Speaker 4: inspire you to finish the song.

127
00:08:00,600 --> 00:08:03,119
Speaker 3: What's your vision of the future, Then at which point

128
00:08:03,160 --> 00:08:06,360
Speaker 3: do you become obsolete or unnecessary?

129
00:08:07,080 --> 00:08:10,680
Speaker 4: If I can imagine a tool that will make and

130
00:08:10,800 --> 00:08:14,080
Speaker 4: generate what you describe more like me Johne for the

131
00:08:14,160 --> 00:08:18,800
Speaker 4: image for example, or Wali, I think orderly, I mean

132
00:08:18,920 --> 00:08:21,480
Speaker 4: I think it will be great, and maybe it will

133
00:08:21,520 --> 00:08:25,440
Speaker 4: replace some musician or it will be game changer.

134
00:08:25,920 --> 00:08:28,200
Speaker 3: The ability to type in a text prompt and get

135
00:08:28,200 --> 00:08:32,600
Speaker 3: a full song back sounds wild and also pretty dangerous

136
00:08:32,600 --> 00:08:35,280
Speaker 3: when you consider those replaced musicians. But let's face it,

137
00:08:35,320 --> 00:08:37,680
Speaker 3: a sense of danger is not going to slow down

138
00:08:37,720 --> 00:08:40,719
Speaker 3: this tech race, and the finish line may be closer

139
00:08:40,760 --> 00:08:44,120
Speaker 3: than we think. Ed Newton Rex is a composer turned

140
00:08:44,120 --> 00:08:46,840
Speaker 3: AI f Sconado. He says this year will be a

141
00:08:47,000 --> 00:08:51,079
Speaker 3: major milestone towards one hundred percent computer written songs, though

142
00:08:51,120 --> 00:08:53,440
Speaker 3: for him, the real value of AI will be in

143
00:08:53,480 --> 00:09:07,280
Speaker 3: collaborating with and inspiring humans. He points to his own

144
00:09:07,320 --> 00:09:09,640
Speaker 3: work I Stand in the Library, which you can hear

145
00:09:09,679 --> 00:09:20,040
Speaker 3: in the background, as an example of that. He wrote

146
00:09:20,080 --> 00:09:36,080
Speaker 3: the music, but he got AI to write the lyrics.

147
00:09:43,600 --> 00:09:45,520
Speaker 7: This isn't just about kind of creating a three minute

148
00:09:45,559 --> 00:09:48,400
Speaker 7: song having AI do that. Once you have an artificial

149
00:09:48,480 --> 00:09:49,640
Speaker 7: entity that's essentially musical.

150
00:09:50,040 --> 00:09:50,920
Speaker 6: There is so much you can do.

151
00:09:50,960 --> 00:09:53,200
Speaker 7: You can ask it for ideas, you can ask it

152
00:09:53,200 --> 00:09:56,240
Speaker 7: for inspiration, you can have interplay with it, you can

153
00:09:56,280 --> 00:09:59,319
Speaker 7: get it to fill in extra parts in the music

154
00:09:59,360 --> 00:10:01,280
Speaker 7: you're making. You can kind of do anything with it

155
00:10:01,320 --> 00:10:03,320
Speaker 7: that you could do with a human musician working with you.

156
00:10:03,559 --> 00:10:05,120
Speaker 7: And I think that's where it gets really exciting.

157
00:10:14,360 --> 00:10:18,360
Speaker 3: All this enthusiasm for AI has kind of surprised me. Clearly,

158
00:10:18,520 --> 00:10:22,000
Speaker 3: Bonoir and Edge showed that there's real human creativity going

159
00:10:22,040 --> 00:10:25,040
Speaker 3: into AI music. I came into this expectant to find

160
00:10:25,080 --> 00:10:30,200
Speaker 3: tone def tech bros, not real musicians. But I'm also

161
00:10:30,280 --> 00:10:34,000
Speaker 3: feeling uneasy about it. Even setting aside the messy ethics

162
00:10:34,120 --> 00:10:37,679
Speaker 3: raised by AI that's trained on other artists' work, I'm

163
00:10:37,720 --> 00:10:41,080
Speaker 3: starting to get nagging flashbacks of the Disney cartoon Fantasia

164
00:10:41,240 --> 00:10:44,959
Speaker 3: I saw as a kid. Remember the Sources Apprentice, when

165
00:10:45,000 --> 00:10:48,640
Speaker 3: Mickey Mouse starts getting addicted to his productive time saving tool,

166
00:10:49,120 --> 00:10:52,440
Speaker 3: creating an army of broomsticks to do his chores chaos.

167
00:10:52,440 --> 00:10:56,640
Speaker 3: Since hughes as control is lost, and imagining an uncontrolled

168
00:10:56,679 --> 00:11:00,480
Speaker 3: release of AI models and voice clones does not inspire confidence.

169
00:11:02,280 --> 00:11:04,920
Speaker 3: I asked the while what the dangers of this might be.

170
00:11:05,600 --> 00:11:08,599
Speaker 4: I think that Eminem remembers all the songs that he

171
00:11:08,679 --> 00:11:12,600
Speaker 4: has done, so he will easily say no, it's not me.

172
00:11:14,360 --> 00:11:17,719
Speaker 4: So I don't think that this kind of demo, this

173
00:11:18,120 --> 00:11:22,600
Speaker 4: kind of process is really dangerous. But hybotization, as I

174
00:11:22,720 --> 00:11:26,120
Speaker 4: show it begins to be dangerous because you can't control.

175
00:11:26,800 --> 00:11:29,920
Speaker 4: It's hard to control. If I didn't said to you

176
00:11:30,120 --> 00:11:33,040
Speaker 4: like like it's I did it with Chet Baker, you

177
00:11:33,400 --> 00:11:37,959
Speaker 4: wouldn't have recognized Chet Baker, So that becomes a bit

178
00:11:38,040 --> 00:11:41,560
Speaker 4: dangerous here. And it's dangerous also because you can do

179
00:11:41,600 --> 00:11:42,360
Speaker 4: it very easily.

180
00:11:45,080 --> 00:11:49,400
Speaker 3: We have to wonder about unattended consequences like misinformation, misuse,

181
00:11:49,440 --> 00:11:53,280
Speaker 3: and bias too. Last year, a fictional computer generated rapper

182
00:11:53,360 --> 00:11:56,560
Speaker 3: called fn Meeker, complete with green hair and nose rings

183
00:11:56,559 --> 00:12:00,000
Speaker 3: and teeth grills, sparked an online backlash after his lyrics

184
00:12:00,160 --> 00:12:04,320
Speaker 3: used racial slurs and trivialized police brutality. Will the toxic

185
00:12:04,360 --> 00:12:07,800
Speaker 3: content we've seen from aichatbots like chat GPT be any

186
00:12:07,800 --> 00:12:11,120
Speaker 3: more acceptable if it's set to music. This is fundamentally

187
00:12:11,120 --> 00:12:16,040
Speaker 3: about humanity versus technology, says musicians, writes advocate Crispin Hunt

188
00:12:16,360 --> 00:12:19,240
Speaker 3: once a member of nineties British pop band The Long Pigs.

189
00:12:19,559 --> 00:12:22,640
Speaker 8: Paul McCartney singing the Beach Boys is really amusing, Or

190
00:12:22,760 --> 00:12:25,400
Speaker 8: Drake and the Weekend doing a duet on somebody else

191
00:12:25,559 --> 00:12:28,920
Speaker 8: is really amusing. I would be totally freaked out if

192
00:12:28,960 --> 00:12:32,760
Speaker 8: somebody took my voice, and I think we as humans

193
00:12:32,840 --> 00:12:36,640
Speaker 8: need to get in front of that very quickly, before

194
00:12:37,280 --> 00:12:40,559
Speaker 8: before the concrete set before it's set in and before

195
00:12:41,240 --> 00:12:45,960
Speaker 8: mistakes get made, or we just give in to this

196
00:12:46,160 --> 00:12:50,040
Speaker 8: idea that you know, oh it's progress, you can't stop progress.

197
00:12:50,120 --> 00:12:54,600
Speaker 8: But there's a great line by a Polish dissident poet,

198
00:12:54,760 --> 00:12:57,840
Speaker 8: courseeddann islav Lek, which said, is it progress for the

199
00:12:57,880 --> 00:12:59,080
Speaker 8: cannibal eats with a fork?

200
00:13:01,040 --> 00:13:01,360
Speaker 2: Ouch?

201
00:13:02,040 --> 00:13:04,400
Speaker 3: So maybe there is a thin line between the well

202
00:13:04,480 --> 00:13:08,760
Speaker 3: meaning digital composers and the evil fork wielding cannibals. The

203
00:13:08,880 --> 00:13:12,080
Speaker 3: next big question is what is the music industry doing

204
00:13:12,120 --> 00:13:15,040
Speaker 3: about it? Well, the first response from the record labels

205
00:13:15,040 --> 00:13:18,559
Speaker 3: has been straight out of the naps to playbook, fight, Fight, Fight.

206
00:13:19,400 --> 00:13:21,960
Speaker 3: But this isn't the year two thousand anymore. They're good

207
00:13:21,960 --> 00:13:25,000
Speaker 3: at it now. When the NP three landed, record companies

208
00:13:25,040 --> 00:13:27,400
Speaker 3: were like a grandparent trying to understand how to work

209
00:13:27,400 --> 00:13:30,720
Speaker 3: their new smartphone. Today, they've rebuilt themselves around a small

210
00:13:30,800 --> 00:13:34,240
Speaker 3: number of streaming platforms like Spotify. It's a lot easier

211
00:13:34,280 --> 00:13:37,120
Speaker 3: to make digital pirates walk the plank, and there's been

212
00:13:37,200 --> 00:13:40,440
Speaker 3: a lot of plank walking. Hearts on My Sleeve was

213
00:13:40,520 --> 00:13:44,160
Speaker 3: quickly taken down from social media. Unauthorized voice cloning apps

214
00:13:44,160 --> 00:13:47,760
Speaker 3: are being shut down, the Grammys have banned fully AI

215
00:13:47,960 --> 00:13:51,760
Speaker 3: generated music, and Spotify is cracking down on bot activity

216
00:13:51,800 --> 00:13:55,920
Speaker 3: on his platform. Here's Jeffrey Halston, general counsel for Universal

217
00:13:56,040 --> 00:13:59,200
Speaker 3: Music Group, whose artists include Taylor Swift and Ed Sheeran,

218
00:13:59,600 --> 00:14:02,559
Speaker 3: testifying before the US Senate Judiciary Committee in July.

219
00:14:03,280 --> 00:14:07,920
Speaker 9: AI in the service of artists and creativity can be

220
00:14:07,960 --> 00:14:12,960
Speaker 9: a very very good thing. But AI that uses, or

221
00:14:13,040 --> 00:14:17,400
Speaker 9: worse yet, appropriates the work of these artists and creators

222
00:14:18,200 --> 00:14:22,480
Speaker 9: in their creative expression, their name, their image, their likeness,

223
00:14:22,680 --> 00:14:28,840
Speaker 9: their voice without authorization, without consent simply is not.

224
00:14:29,320 --> 00:14:29,880
Speaker 5: A good thing.

225
00:14:30,680 --> 00:14:33,360
Speaker 3: And the music industry has also learned other lessons since

226
00:14:33,400 --> 00:14:36,120
Speaker 3: the napster days. It knows if people really want to

227
00:14:36,240 --> 00:14:39,480
Speaker 3: use something, death by lawsuit is not gonna work, and

228
00:14:39,600 --> 00:14:42,600
Speaker 3: not all AI music is deep fate. Users of AI

229
00:14:42,640 --> 00:14:47,160
Speaker 3: website Boomy have created over seventeen million weird sounding electronic tracks.

230
00:14:47,160 --> 00:14:49,960
Speaker 3: For example, I have trouble getting through just one of them.

231
00:14:50,000 --> 00:14:52,480
Speaker 3: But still maybe one day there'll be a hit in there,

232
00:14:52,760 --> 00:14:54,760
Speaker 3: and not one that can be shut down by a lawyer.

233
00:14:55,280 --> 00:14:59,120
Speaker 3: So here's the next twist. Record labels are done just

234
00:14:59,200 --> 00:15:02,480
Speaker 3: playing defense. They now want to go on the offensive

235
00:15:02,800 --> 00:15:06,240
Speaker 3: by making some AI music of their own that plays

236
00:15:06,240 --> 00:15:07,080
Speaker 3: by their rules.

237
00:15:07,200 --> 00:15:09,600
Speaker 2: Wait, wait, hold it right there, linell. This is shaping

238
00:15:09,680 --> 00:15:11,480
Speaker 2: up to be quite the story, but we need to

239
00:15:11,480 --> 00:15:13,640
Speaker 2: take a quick break and I'll see you on the

240
00:15:13,640 --> 00:15:14,120
Speaker 2: other side.

241
00:15:14,480 --> 00:15:15,120
Speaker 3: Sounds good to me.

242
00:15:20,880 --> 00:15:24,160
Speaker 2: We're back with Lionell Lurn talking about AI and music.

243
00:15:25,240 --> 00:15:28,160
Speaker 2: So Lionelle, let's hear more about how record labels are

244
00:15:28,160 --> 00:15:30,320
Speaker 2: fighting back by co opting AI.

245
00:15:31,360 --> 00:15:34,680
Speaker 3: Sure, it's a story that begins with a young entrepreneur

246
00:15:34,800 --> 00:15:38,560
Speaker 3: based in Berlin called Oleg Stovitski, the AI starts up.

247
00:15:38,640 --> 00:15:42,280
Speaker 3: He co founded Endel, made the music that you're listening

248
00:15:42,320 --> 00:15:52,600
Speaker 3: to now. And Endel has just signed a landmark partnership

249
00:15:52,880 --> 00:15:56,800
Speaker 3: with Universal Music group, that really anti deep fake label

250
00:15:56,840 --> 00:16:00,320
Speaker 3: we heard from earlier, to make AI music use their

251
00:16:00,400 --> 00:16:03,760
Speaker 3: huge catogy stars. And this all started with Oleg looking

252
00:16:03,760 --> 00:16:07,160
Speaker 3: for music that would help them concentrate, similar to the

253
00:16:07,200 --> 00:16:09,680
Speaker 3: work of ambient music pioneer Brian Eno.

254
00:16:10,440 --> 00:16:13,160
Speaker 5: Every time I'm trying to concentrate, I listened to Brain,

255
00:16:13,240 --> 00:16:16,240
Speaker 5: you know, and then noticed that a lot of his music,

256
00:16:16,240 --> 00:16:19,080
Speaker 5: and a lot of ambient music in general started kind

257
00:16:19,120 --> 00:16:22,640
Speaker 5: of popping up. When all of these functional music playlists

258
00:16:22,680 --> 00:16:26,000
Speaker 5: and projects like music for Programming appeared.

259
00:16:26,520 --> 00:16:29,800
Speaker 3: That got Oleg thinking, what if every time you hit play,

260
00:16:30,240 --> 00:16:33,680
Speaker 3: the soundscape was different and fit your mood by monitoring

261
00:16:33,760 --> 00:16:36,880
Speaker 3: data such as your real time heart rate or the

262
00:16:36,880 --> 00:16:39,400
Speaker 3: time of day. That would be like having your own

263
00:16:39,480 --> 00:16:42,040
Speaker 3: personal Brian Eno, made by a machine.

264
00:16:42,640 --> 00:16:44,320
Speaker 5: I kind of turned to my co founders and I

265
00:16:44,360 --> 00:16:46,880
Speaker 5: was like, well, I guess we're going to have to

266
00:16:46,880 --> 00:16:47,680
Speaker 5: build an AI.

267
00:16:48,360 --> 00:16:51,400
Speaker 3: This AI would be trained on a very specific set

268
00:16:51,400 --> 00:16:55,120
Speaker 3: of musical building blocks, or stems. Stems can be thought

269
00:16:55,120 --> 00:16:58,360
Speaker 3: of as these separate instruments or components that make up

270
00:16:58,360 --> 00:17:04,879
Speaker 3: a song. Think of an individual drum beet, a keyboard line,

271
00:17:07,400 --> 00:17:08,320
Speaker 3: a vocal melody.

272
00:17:08,840 --> 00:17:12,720
Speaker 2: I've been to the top of Mount Everest, I've.

273
00:17:12,560 --> 00:17:17,880
Speaker 9: Sailed the seven seats, I've shared the stage with all

274
00:17:17,960 --> 00:17:19,239
Speaker 9: the best.

275
00:17:20,119 --> 00:17:25,199
Speaker 3: As three separate stems, Endel's AI will then uniquely reassemble

276
00:17:25,240 --> 00:17:29,160
Speaker 3: the stems into a new soundscape every time. Personally, it's

277
00:17:29,160 --> 00:17:31,399
Speaker 3: not for me, but clearly there is a market for

278
00:17:31,440 --> 00:17:34,520
Speaker 3: functional music. It represents an estimated seven to ten percent

279
00:17:34,640 --> 00:17:37,879
Speaker 3: of the entire streaming market. Endel's next smart move was

280
00:17:37,880 --> 00:17:41,600
Speaker 3: to blend that robotic AI with well known human artists

281
00:17:41,680 --> 00:17:45,919
Speaker 3: like James Blake and Grimes. They submitted the stems and

282
00:17:46,119 --> 00:17:49,160
Speaker 3: Endal did the rest and basic. This all on stems

283
00:17:49,520 --> 00:17:53,840
Speaker 3: rather than existing albums or songs, means Endel gets official

284
00:17:53,880 --> 00:17:58,640
Speaker 3: recognition as co composer of the real time soundscape. That's

285
00:17:58,680 --> 00:18:02,160
Speaker 3: how Oleg says he avoided a fight with the music industry.

286
00:18:02,680 --> 00:18:06,080
Speaker 5: With us, it was always the reaction was always a

287
00:18:06,160 --> 00:18:12,159
Speaker 5: mix of through kind of genuine interest and almost excitement,

288
00:18:12,359 --> 00:18:15,960
Speaker 5: and people were intrigued because we never removed the artists

289
00:18:16,000 --> 00:18:18,480
Speaker 5: from this process. We never said, oh, you know, we

290
00:18:18,600 --> 00:18:20,600
Speaker 5: just get the stamps from James Blake, but then we're

291
00:18:20,640 --> 00:18:23,919
Speaker 5: going to produce like a million records and James is

292
00:18:23,960 --> 00:18:25,919
Speaker 5: not going to be mentioned anymore.

293
00:18:26,680 --> 00:18:30,480
Speaker 3: But the bigger hand'ed go, the more competition became unavoidable

294
00:18:30,760 --> 00:18:32,800
Speaker 3: because money that would normally be going to the record

295
00:18:32,840 --> 00:18:36,879
Speaker 3: labels is being diverted to text us ups making ambient music.

296
00:18:37,560 --> 00:18:40,080
Speaker 3: The economics of streaming today are all about market share.

297
00:18:40,480 --> 00:18:43,080
Speaker 3: Listen to a track on Spotify for more than thirty seconds,

298
00:18:43,280 --> 00:18:46,440
Speaker 3: and it counts as a play, whether it's forty seconds

299
00:18:46,440 --> 00:19:00,280
Speaker 3: of experimental electronic music, a three minute pop hit, five

300
00:19:00,320 --> 00:19:08,160
Speaker 3: minutes of whale song. The more people click on ambient

301
00:19:08,240 --> 00:19:12,480
Speaker 3: music rather than say, drake, the more it'll impact the

302
00:19:12,480 --> 00:19:16,440
Speaker 3: bottom line. According to Golden Sachs, the three major labels

303
00:19:16,520 --> 00:19:20,199
Speaker 3: market share has dropped from around eighty five percent to

304
00:19:20,280 --> 00:19:24,520
Speaker 3: seventy one percent between twenty seventeen and twenty twenty two,

305
00:19:24,600 --> 00:19:27,960
Speaker 3: which is why when the announcement of that landmark universal

306
00:19:28,080 --> 00:19:32,320
Speaker 3: Endel deal for more record label approved AI music landed

307
00:19:32,359 --> 00:19:35,080
Speaker 3: in May, it looks a lot like can't beat them,

308
00:19:35,560 --> 00:19:36,439
Speaker 3: join them?

309
00:19:36,680 --> 00:19:39,000
Speaker 5: You know they can. I just keep ordering takedowns. That

310
00:19:39,200 --> 00:19:42,560
Speaker 5: is a road to nowhere. They need to have kind

311
00:19:42,600 --> 00:19:46,879
Speaker 5: of an offensive strategy and they need to embrace this

312
00:19:46,960 --> 00:19:50,280
Speaker 5: in some way in form and it's just turned out

313
00:19:50,280 --> 00:19:52,480
Speaker 5: to be that Endel was kind of the only company

314
00:19:52,600 --> 00:19:55,800
Speaker 5: that spent years kind of learning to speak the language,

315
00:19:55,800 --> 00:19:57,760
Speaker 5: and then it's the dance of these labels.

316
00:19:58,480 --> 00:20:03,040
Speaker 3: So what does a record label backed ambient AI do next?

317
00:20:07,880 --> 00:20:11,920
Speaker 3: Ole extreme is to make new AI versions of artists

318
00:20:12,000 --> 00:20:16,320
Speaker 3: existing catalogs. So, for example, Ed Sheeran or Taylor Swift

319
00:20:16,800 --> 00:20:20,320
Speaker 3: might one day offer Endel versions of their albums designed

320
00:20:20,320 --> 00:20:24,800
Speaker 3: for sleep, work or exercise. And imagine if in the future,

321
00:20:24,880 --> 00:20:29,200
Speaker 3: instead of Taylor's version, we get happy, sad or heavy

322
00:20:29,200 --> 00:20:30,680
Speaker 3: metal versions of her albums.

323
00:20:31,119 --> 00:20:33,919
Speaker 5: Essentially expands the universe of the artists. And it's just

324
00:20:33,960 --> 00:20:39,439
Speaker 5: such a beautiful concept of creating a functional soundscape version

325
00:20:40,160 --> 00:20:41,440
Speaker 5: of your favorite artist music.

326
00:20:41,760 --> 00:20:45,160
Speaker 3: It's not just Endel. Other labels are sensing an opportunity.

327
00:20:45,680 --> 00:20:49,560
Speaker 3: Warner Music's CEO recently praised an AI enabled duet between

328
00:20:49,600 --> 00:20:53,280
Speaker 3: Costa Rican musician Pedro Capmani and the generated voice of

329
00:20:53,280 --> 00:21:01,600
Speaker 3: his late father. Oh I asked Gairi, head of music

330
00:21:01,600 --> 00:21:05,240
Speaker 3: group Believe, which owns a DIY music platform called tune Call,

331
00:21:05,760 --> 00:21:08,840
Speaker 3: about this change of heart. Some might say double think

332
00:21:09,000 --> 00:21:10,240
Speaker 3: in the industry.

333
00:21:10,280 --> 00:21:13,520
Speaker 10: Whether it's for Believe or generally for the music industry,

334
00:21:15,080 --> 00:21:18,560
Speaker 10: is an opportunity and not a threat. Why do people

335
00:21:18,680 --> 00:21:23,600
Speaker 10: think AI is a threat generally because they do think

336
00:21:23,600 --> 00:21:26,520
Speaker 10: that the deployment of AI will not be controlled and

337
00:21:26,600 --> 00:21:30,600
Speaker 10: af the opposite view around this, which is that AA

338
00:21:30,680 --> 00:21:36,040
Speaker 10: deployment will be heavily controlled and everyone is approaching AI

339
00:21:36,640 --> 00:21:39,760
Speaker 10: with very a very strong sense of responsibility.

340
00:21:39,920 --> 00:21:42,960
Speaker 3: That control means labels can set their own rules of

341
00:21:42,960 --> 00:21:46,760
Speaker 3: engagement with AI. Deri gives four examples of rules that

342
00:21:46,800 --> 00:21:50,640
Speaker 3: Believe is currently applying on AI. Music uploaded to its platform.

343
00:21:50,920 --> 00:21:56,240
Speaker 3: Consent from the artist, control over the output, compensation, and transparency.

344
00:21:56,440 --> 00:22:00,119
Speaker 10: That's how we're approaching the landscape today, and so we

345
00:22:00,200 --> 00:22:03,639
Speaker 10: are not going to distribute tracks that are one hundred

346
00:22:03,680 --> 00:22:08,000
Speaker 10: percent created by AI unless they meet some of these

347
00:22:08,040 --> 00:22:08,800
Speaker 10: grit areas.

348
00:22:09,280 --> 00:22:12,400
Speaker 3: Believe operates mainly in the mid market of artists rather

349
00:22:12,400 --> 00:22:15,439
Speaker 3: than the superstars, and the hope is that industry friendly

350
00:22:15,480 --> 00:22:18,760
Speaker 3: AI can dig up the next generation's bright young things

351
00:22:19,200 --> 00:22:21,800
Speaker 3: by making it easier to unlock tunes sitting inside their

352
00:22:21,800 --> 00:22:24,879
Speaker 3: heads again or while following the rules.

353
00:22:25,240 --> 00:22:28,399
Speaker 10: We think the number of artists that gets educated for

354
00:22:28,560 --> 00:22:32,159
Speaker 10: technology to become better musician is going to increase, and

355
00:22:32,560 --> 00:22:36,159
Speaker 10: more people creating better music will find an audience. This

356
00:22:36,320 --> 00:22:40,679
Speaker 10: is going to help frontline artists elevate and create better music.

357
00:22:41,000 --> 00:22:44,359
Speaker 3: This optimism around control is also why YouTube has created

358
00:22:44,400 --> 00:22:47,160
Speaker 3: a new set of rules for what it calls responsible

359
00:22:47,200 --> 00:22:50,719
Speaker 3: AI music. But this all seems to be moving very fast.

360
00:22:50,960 --> 00:22:52,879
Speaker 3: There's a lot of pressure on labels to be a

361
00:22:52,920 --> 00:22:56,520
Speaker 3: first mover and sign deals with tech companies. I'll be

362
00:22:56,600 --> 00:23:00,160
Speaker 3: sure that artists consent and all that precious data from

363
00:23:00,160 --> 00:23:03,639
Speaker 3: their life's work isn't being handed over a little too quickly.

364
00:23:04,440 --> 00:23:08,280
Speaker 3: And what if the focus on control starts to stifle creativity?

365
00:23:08,560 --> 00:23:11,320
Speaker 3: If it trips up artists who are using AI but

366
00:23:11,440 --> 00:23:16,280
Speaker 3: fairly without breaching copyright. Remember ed Newton Rex, who he

367
00:23:16,280 --> 00:23:19,000
Speaker 3: heard from earlier in the show. In twenty fifteen, he

368
00:23:19,080 --> 00:23:22,960
Speaker 3: wrapped about this exact thing when promoting his startup jute Deck,

369
00:23:23,640 --> 00:23:27,080
Speaker 3: citing the copyright constraints faced by YouTubers looking for good

370
00:23:27,080 --> 00:23:27,840
Speaker 3: background music.

371
00:23:28,560 --> 00:23:31,000
Speaker 6: Imagine there's a video that you're creating, it's been weeks

372
00:23:31,000 --> 00:23:33,119
Speaker 6: in the making, and your painstaking need take in it

373
00:23:33,160 --> 00:23:35,200
Speaker 6: praying by prayer, have got the whole take looking really

374
00:23:35,200 --> 00:23:37,960
Speaker 6: super great. Start to find the right music. If you've

375
00:23:37,960 --> 00:23:39,959
Speaker 6: got a song you like, you're not allowed to use it.

376
00:23:40,160 --> 00:23:42,920
Speaker 6: Copyright means if you choose it for your video, YouTube

377
00:23:43,000 --> 00:23:45,920
Speaker 6: will just remove it. But this right stuff that's all unique.

378
00:23:45,920 --> 00:23:47,640
Speaker 6: It will do it in any style U.

379
00:23:47,600 --> 00:23:53,040
Speaker 3: See come Back, Fake Drake, all is forgiven? And going

380
00:23:53,119 --> 00:23:56,520
Speaker 3: back to the Sorcerer's Apprentice, what if record label robots

381
00:23:56,560 --> 00:24:00,199
Speaker 3: become so effective at expanding their catalog of stars that

382
00:24:00,280 --> 00:24:03,040
Speaker 3: it makes it harder for the new generation of artists

383
00:24:03,080 --> 00:24:06,280
Speaker 3: to cut through the noise. When does the revolution become stagnation.

384
00:24:07,280 --> 00:24:09,760
Speaker 3: It's time to ask if there's an alternative to either

385
00:24:09,800 --> 00:24:13,560
Speaker 3: the dystopia of no control or the utopia of total control,

386
00:24:14,280 --> 00:24:16,880
Speaker 3: something like a third way that's maybe a little more

387
00:24:16,880 --> 00:24:20,440
Speaker 3: punk rock, putting more power into the hands of the artists.

388
00:24:20,680 --> 00:24:23,600
Speaker 3: And the best person to ask is Grimes, who we

389
00:24:23,640 --> 00:24:25,560
Speaker 3: heard from at the beginning of the show. She is

390
00:24:25,640 --> 00:24:30,720
Speaker 3: one famous artist who is actually embracing AI experimentation while

391
00:24:30,760 --> 00:24:33,639
Speaker 3: also still trying to set limits, sort of like rapper

392
00:24:33,720 --> 00:24:37,399
Speaker 3: Chuck Dee, who during the Napster era said downloadable music

393
00:24:37,520 --> 00:24:41,080
Speaker 3: was a model to embrace. I asked Grimes first off,

394
00:24:41,119 --> 00:24:44,440
Speaker 3: if she generally felt excited or worried about the current

395
00:24:44,440 --> 00:24:45,600
Speaker 3: state of AI and music.

396
00:24:45,840 --> 00:24:47,920
Speaker 1: I think I would say I'm quite both. One of

397
00:24:47,960 --> 00:24:49,600
Speaker 1: the main things I'm doing right now is trying to

398
00:24:49,640 --> 00:24:52,600
Speaker 1: figure out the ways to make this safe for humans,

399
00:24:52,880 --> 00:24:56,920
Speaker 1: not even just existentially, but like socially and emotionally and civilizationally.

400
00:24:57,080 --> 00:24:58,760
Speaker 1: But I also think that's totally possible.

401
00:24:59,080 --> 00:25:02,040
Speaker 3: One of her biggest experiments is a new project allowing

402
00:25:02,080 --> 00:25:05,080
Speaker 3: deep fakes or songs that clone her voice, which she

403
00:25:05,119 --> 00:25:08,280
Speaker 3: announced when Hearts of My Sleeve went viral. She tweeted

404
00:25:08,320 --> 00:25:11,840
Speaker 3: in April, feel free to use my voice without penalty.

405
00:25:12,160 --> 00:25:15,520
Speaker 3: I have no label and no legal bindings. I guess

406
00:25:15,520 --> 00:25:19,160
Speaker 3: it's not surprising Grimes as into AI, her dark electronic

407
00:25:19,200 --> 00:25:22,679
Speaker 3: music already embodies a kind of tech human coexistence. She

408
00:25:22,720 --> 00:25:27,000
Speaker 3: cultivates an extremely online sci fi anime persona, live streams

409
00:25:27,040 --> 00:25:30,560
Speaker 3: video games, and assault digital art as NFTs. When I

410
00:25:30,600 --> 00:25:32,840
Speaker 3: asked her how she sees AI being used in music,

411
00:25:33,119 --> 00:25:36,399
Speaker 3: she compares it to auto tune, that instantly recognizable tool

412
00:25:36,480 --> 00:25:38,240
Speaker 3: allowing anyone to hit those high notes.

413
00:25:38,760 --> 00:25:43,520
Speaker 1: I am really pro democratizing the process. I have a

414
00:25:43,520 --> 00:25:46,440
Speaker 1: pretty interesting, recognizable timber to my voice, but I have

415
00:25:46,520 --> 00:25:48,679
Speaker 1: a weird speech, like ever Lis been, Like I just

416
00:25:48,720 --> 00:25:50,720
Speaker 1: am not that in tune and I don't really care

417
00:25:50,760 --> 00:25:53,280
Speaker 1: about being in tune, and like a producer, autotune is

418
00:25:53,280 --> 00:25:55,800
Speaker 1: a good sort of like metaphor, because I think we got,

419
00:25:55,920 --> 00:25:58,399
Speaker 1: especially in the last decade a lot more interesting and

420
00:25:58,480 --> 00:26:01,600
Speaker 1: unique top blind than we'd gotten previous to that, because

421
00:26:01,600 --> 00:26:03,800
Speaker 1: you had all these kind of brains that are not

422
00:26:03,880 --> 00:26:07,720
Speaker 1: singer brains, like singing really well. And I think you

423
00:26:07,800 --> 00:26:09,760
Speaker 1: especially get magic a lot of the time when you

424
00:26:09,800 --> 00:26:12,040
Speaker 1: have someone who doesn't have the fundamental or it wasn't

425
00:26:12,080 --> 00:26:14,159
Speaker 1: super trained in something and they just try it for

426
00:26:14,160 --> 00:26:14,800
Speaker 1: the first time.

427
00:26:15,280 --> 00:26:18,480
Speaker 3: So the more AI can support creators by lowering the

428
00:26:18,480 --> 00:26:22,320
Speaker 3: barriers to entry for composition and production, the better that

429
00:26:22,400 --> 00:26:28,679
Speaker 3: melody you just wrote on an acoustic guitar could quickly

430
00:26:28,720 --> 00:26:31,399
Speaker 3: sound like a full band in the studio without actually

431
00:26:31,400 --> 00:26:32,960
Speaker 3: needing the band or the studio.

432
00:26:33,200 --> 00:26:35,119
Speaker 1: And I personally know a lot of artists who have

433
00:26:35,160 --> 00:26:37,800
Speaker 1: really struggled the process of finding people who are good

434
00:26:37,840 --> 00:26:40,679
Speaker 1: and then paying them can be really hard, and especially

435
00:26:40,760 --> 00:26:44,080
Speaker 1: with the complexity of music publishing, that can be really hard.

436
00:26:44,119 --> 00:26:45,760
Speaker 1: So if you can just sing a song and then

437
00:26:45,800 --> 00:26:49,120
Speaker 1: get like a bunch of different generations of instrumentals behind it,

438
00:26:49,680 --> 00:26:52,760
Speaker 1: I think for singers specifically and songwriters, there's a lot

439
00:26:52,800 --> 00:26:53,680
Speaker 1: of opportunity there.

440
00:26:53,800 --> 00:26:57,600
Speaker 3: Grime's inspiration for her own initial experiments with AI came

441
00:26:57,640 --> 00:26:59,800
Speaker 3: from the idea of training a model on her own

442
00:26:59,800 --> 00:27:03,840
Speaker 3: work bottling her entire style into a product, not just

443
00:27:03,920 --> 00:27:08,200
Speaker 3: one particular sample, effect or tone. She imagined a database

444
00:27:08,240 --> 00:27:11,679
Speaker 3: of all sorts of different music producers from Timberland to

445
00:27:11,960 --> 00:27:15,359
Speaker 3: underground hadent like the late producer Sophie, getting a reliable

446
00:27:15,359 --> 00:27:18,440
Speaker 3: income from selling this type of data. And then came

447
00:27:18,480 --> 00:27:21,320
Speaker 3: the hype and the panic around Hard on my Sleeve,

448
00:27:21,880 --> 00:27:24,159
Speaker 3: which served as a kind of wake up call to

449
00:27:24,160 --> 00:27:25,680
Speaker 3: get the debate around AI going.

450
00:27:26,040 --> 00:27:29,320
Speaker 1: I think it's good for people to push the limits

451
00:27:29,480 --> 00:27:32,720
Speaker 1: with what's legal so everyone sees what the technology can do.

452
00:27:32,920 --> 00:27:34,680
Speaker 1: I tried to make my voice a bunch of times,

453
00:27:34,680 --> 00:27:36,200
Speaker 1: but the technology just hadn't been there.

454
00:27:36,920 --> 00:27:41,080
Speaker 3: Grimes partnered with Create Safe to launch software called Elftech,

455
00:27:41,280 --> 00:27:44,439
Speaker 3: which uses AI to replicate her voice. The way it

456
00:27:44,440 --> 00:27:47,879
Speaker 3: works is this music creators can upload audio or record

457
00:27:47,880 --> 00:27:50,600
Speaker 3: directly into the app, then receive a file with the

458
00:27:50,600 --> 00:27:53,320
Speaker 3: same audio, but with Grinder's voice instead of their own.

459
00:27:53,480 --> 00:27:56,119
Speaker 3: If the song gets an official release, the royalties are

460
00:27:56,160 --> 00:28:00,480
Speaker 3: split fifty to fifty. Grinder's manager Lenard Doda told me

461
00:28:00,560 --> 00:28:02,960
Speaker 3: when we spoke that fifty thousand people are using the

462
00:28:03,040 --> 00:28:06,240
Speaker 3: product and they've had over one hundred thousand audio generations

463
00:28:06,720 --> 00:28:09,560
Speaker 3: uploading their audio or playing with the tool. Over the

464
00:28:09,600 --> 00:28:11,560
Speaker 3: more than one hundred grimes Ai songs that have been

465
00:28:11,600 --> 00:28:15,359
Speaker 3: put on streaming platforms, one has over six hundred thousand plays,

466
00:28:15,359 --> 00:28:18,159
Speaker 3: he says. And if you're wondering what grimes clone actually

467
00:28:18,160 --> 00:28:21,960
Speaker 3: sounds like, here's a track by Bnoit Care Yes, mister

468
00:28:22,000 --> 00:28:25,720
Speaker 3: Daddy's car himself called us Noir, in which grimes Ai

469
00:28:26,080 --> 00:28:27,800
Speaker 3: replaces his vocals sung in.

470
00:28:27,760 --> 00:28:43,120
Speaker 1: French Santan palin coofen Hesiod custom O Gas.

471
00:28:45,120 --> 00:28:48,520
Speaker 3: Compared to how collaborations are normally done, this is way

472
00:28:48,640 --> 00:28:54,080
Speaker 3: less paperwork, way less hassle, and way more streamlined than usual.

473
00:28:54,720 --> 00:28:59,080
Speaker 3: Now there isn't just one Grimes, but potentially infinite Grimes,

474
00:28:59,440 --> 00:29:02,080
Speaker 3: which is and always online twenty four to seven world

475
00:29:02,160 --> 00:29:05,440
Speaker 3: of social media and music streaming is a huge advantage.

476
00:29:10,680 --> 00:29:14,280
Speaker 3: This takes fan engagement into a new era. Grimes also

477
00:29:14,400 --> 00:29:17,200
Speaker 3: reckons the music being created in the more AI future

478
00:29:17,520 --> 00:29:20,200
Speaker 3: could actually sound more exciting than today's all you can

479
00:29:20,280 --> 00:29:21,719
Speaker 3: listen to streaming buffet.

480
00:29:22,320 --> 00:29:25,640
Speaker 1: If generative music is deployed very carefully and safely, you

481
00:29:25,720 --> 00:29:28,360
Speaker 1: still create a environment where people want to create. You

482
00:29:28,480 --> 00:29:30,840
Speaker 1: incentivize It's like you want to be a really unique

483
00:29:30,880 --> 00:29:33,640
Speaker 1: producer so that you have a unique sound that it's

484
00:29:33,760 --> 00:29:36,720
Speaker 1: worth training a model off of that is like recognizable.

485
00:29:37,040 --> 00:29:39,360
Speaker 1: I feel like a lot of times the algorithms, like

486
00:29:39,400 --> 00:29:42,200
Speaker 1: the streaming algorithms, really encourage music to sound the same.

487
00:29:42,640 --> 00:29:44,640
Speaker 1: This is something that I think would really encourage music

488
00:29:44,680 --> 00:29:46,880
Speaker 1: to sound more different. And I think that is like

489
00:29:46,960 --> 00:29:47,520
Speaker 1: a net good.

490
00:29:48,040 --> 00:29:51,280
Speaker 3: So far, so good. But Grimes is not a utopian.

491
00:29:51,440 --> 00:29:54,480
Speaker 3: She's also aware of the risks. I asked her what

492
00:29:54,640 --> 00:29:57,200
Speaker 3: her red lines are, what the downsides are to this.

493
00:29:57,800 --> 00:29:59,840
Speaker 1: I think, especially if the deep fakes, like people should

494
00:29:59,880 --> 00:30:03,720
Speaker 1: have consent. I think that's super important, super essential, Like

495
00:30:03,760 --> 00:30:06,040
Speaker 1: even though I'm doing this with my voice, I think

496
00:30:06,080 --> 00:30:09,120
Speaker 1: there is potentially an argument that deefix should be illegal

497
00:30:09,240 --> 00:30:14,680
Speaker 1: across the board. I don't actually see a long term

498
00:30:14,800 --> 00:30:16,560
Speaker 1: utility for deep fix, like I see a lot of

499
00:30:16,600 --> 00:30:21,960
Speaker 1: potential for political distress and like political unrest and harming

500
00:30:23,040 --> 00:30:25,920
Speaker 1: especially women. Is it worth the emotional downsides? I don't know.

501
00:30:26,000 --> 00:30:28,040
Speaker 1: That's part why I'm doing this, is running the experiment.

502
00:30:28,760 --> 00:30:32,560
Speaker 3: Grimes has a point. It's tremendously powerful to imagine artists

503
00:30:33,240 --> 00:30:36,280
Speaker 3: rather than tech platforms or record labels, leading the charge

504
00:30:36,320 --> 00:30:39,040
Speaker 3: of AI clones for their own benefit, But the issue

505
00:30:39,080 --> 00:30:43,160
Speaker 3: of control is still there. What if malicious actors, hackers,

506
00:30:43,560 --> 00:30:47,480
Speaker 3: fraudsters use Grimes's voice to sing about hate crimes or

507
00:30:47,560 --> 00:30:50,320
Speaker 3: to ruin her image. What if her own work becomes

508
00:30:50,360 --> 00:30:52,720
Speaker 3: devalued as a result. What if a lot of bad

509
00:30:52,800 --> 00:30:55,880
Speaker 3: content gets uploaded. If there's one big challenge hanging over

510
00:30:55,920 --> 00:30:59,120
Speaker 3: a platform like elf tech, it's probably long term sustainability.

511
00:30:59,400 --> 00:31:02,040
Speaker 3: It takes for alsays in bandwidth to believe a platform.

512
00:31:02,440 --> 00:31:04,760
Speaker 3: So why do it makes sense that models would be

513
00:31:04,840 --> 00:31:08,320
Speaker 3: run at the individual artist level. It should be done

514
00:31:08,360 --> 00:31:12,040
Speaker 3: with a pragmatic, not idealistic approach. This is all moving

515
00:31:12,200 --> 00:31:14,560
Speaker 3: very fast, and Grime says we need to slow down

516
00:31:14,920 --> 00:31:17,040
Speaker 3: and think of all sorts of ways to minimize AI

517
00:31:17,160 --> 00:31:18,720
Speaker 3: harms and maximize benefits.

518
00:31:19,200 --> 00:31:21,440
Speaker 1: I think a lot of people think it's just going

519
00:31:21,480 --> 00:31:23,400
Speaker 1: to be great. Before we just make these things, They're

520
00:31:23,440 --> 00:31:24,880
Speaker 1: just going to be great and we're fine. And a

521
00:31:24,920 --> 00:31:27,440
Speaker 1: lot of people think this is definitely a dystopia and

522
00:31:27,480 --> 00:31:30,760
Speaker 1: we're definitely fund and I think it can truly be

523
00:31:31,000 --> 00:31:34,600
Speaker 1: a third thing. It just is going to require a

524
00:31:34,720 --> 00:31:37,320
Speaker 1: lot of working together and stuff that we have not

525
00:31:37,440 --> 00:31:40,040
Speaker 1: always displayed. We haven't always been great at that, but

526
00:31:40,360 --> 00:31:43,120
Speaker 1: I really truly think there's no reason why we couldn't

527
00:31:43,240 --> 00:31:43,800
Speaker 1: be better at that.

528
00:31:43,880 --> 00:31:45,920
Speaker 5: Now, what's your vision of a future?

529
00:31:46,040 --> 00:31:49,160
Speaker 3: How we consume music? And where is your voice? Is

530
00:31:49,160 --> 00:31:51,920
Speaker 3: it everywhere? What's your what's your for me?

531
00:31:52,120 --> 00:31:56,120
Speaker 1: Personally? I love that, I love I always use these

532
00:31:56,120 --> 00:31:58,440
Speaker 1: two examples. But I think League of Legends and Harry

533
00:31:58,520 --> 00:32:01,360
Speaker 1: Potter have done a very good job of not punishing

534
00:32:01,520 --> 00:32:04,800
Speaker 1: fan art. They don't seem to issue takedowns like they don't.

535
00:32:05,400 --> 00:32:07,600
Speaker 1: They let people sell prints of like League of Legends

536
00:32:07,680 --> 00:32:10,680
Speaker 1: characters and stuff, right people, There's just like it issanly

537
00:32:10,720 --> 00:32:13,400
Speaker 1: prolific amount of Harry Potter fan art, and it makes

538
00:32:13,880 --> 00:32:20,040
Speaker 1: the community richer and super engaged. And Grimes is like

539
00:32:20,160 --> 00:32:22,640
Speaker 1: pretty different from how I am and I see rhymes.

540
00:32:22,800 --> 00:32:24,800
Speaker 1: It's more of a character as a science fiction project.

541
00:32:24,880 --> 00:32:27,080
Speaker 1: It's just all these things, And I don't feel bad

542
00:32:27,200 --> 00:32:30,440
Speaker 1: about not having ownership of that. I would like it

543
00:32:30,640 --> 00:32:34,440
Speaker 1: to not do harmful things. But it's like you look

544
00:32:34,480 --> 00:32:38,560
Speaker 1: at like Disney, It's like a lot of the best

545
00:32:38,600 --> 00:32:40,640
Speaker 1: Disney stuff was made after Whalt Disney was dead.

546
00:32:41,320 --> 00:32:44,320
Speaker 3: So maybe there is that third way between Utopia and

547
00:32:44,400 --> 00:32:47,120
Speaker 3: dystopia after all, And maybe there is a chance at

548
00:32:47,120 --> 00:32:49,880
Speaker 3: a more punk rock independent approach and what the big

549
00:32:49,920 --> 00:32:52,600
Speaker 3: companies are offering. But we soon need to talk about

550
00:32:52,640 --> 00:32:56,120
Speaker 3: some of the real existential risks that are being posed

551
00:32:56,440 --> 00:33:00,080
Speaker 3: by AI, and also why Rick Astley could be the

552
00:33:00,160 --> 00:33:03,800
Speaker 3: key to one of the most burning issues out there, copyright.

553
00:33:04,400 --> 00:33:06,000
Speaker 2: Did you say, Rick Eshlee?

554
00:33:06,520 --> 00:33:08,800
Speaker 3: Yes, Tim, I'm never going to give you up, never

555
00:33:08,880 --> 00:33:11,320
Speaker 3: going to let you down. And actually that's the closest

556
00:33:11,320 --> 00:33:12,680
Speaker 3: we're ever going to get to a Rick role.

557
00:33:14,240 --> 00:33:17,360
Speaker 2: Okay, I'm excited to hear more as I've been throughout

558
00:33:17,400 --> 00:33:20,479
Speaker 2: this episode, but let's take another quick break and then

559
00:33:20,600 --> 00:33:28,040
Speaker 2: pick this back up on the other side. Sure, Okay,

560
00:33:28,240 --> 00:33:31,600
Speaker 2: we're back with Bloomberg opinion columnist Lionel Laurent, and we're

561
00:33:31,640 --> 00:33:33,640
Speaker 2: reaching the end of our journey into the world of

562
00:33:33,760 --> 00:33:36,640
Speaker 2: AI and music. What's next, Lionel?

563
00:33:40,080 --> 00:33:42,360
Speaker 3: So we've talked a lot about them big techtonic shifts

564
00:33:42,360 --> 00:33:45,360
Speaker 3: around our music, but we haven't really talked about the

565
00:33:45,480 --> 00:33:48,719
Speaker 3: existential threat at the heart of this story. There are

566
00:33:48,800 --> 00:33:51,680
Speaker 3: two hundred and thirteen thousand, seven hundred and thirty eight

567
00:33:51,720 --> 00:33:54,840
Speaker 3: identified singers and musicians out there, according to the twenty

568
00:33:54,880 --> 00:33:58,240
Speaker 3: sixteen US Census, and that is, of course only talking

569
00:33:58,280 --> 00:34:00,800
Speaker 3: about the States, they won't be able to keep up

570
00:34:00,800 --> 00:34:04,240
Speaker 3: with a deep pocketed AI industry that's attracting money and talent.

571
00:34:04,920 --> 00:34:07,440
Speaker 3: And even if we keep telling ourselves we'll always value

572
00:34:07,520 --> 00:34:11,359
Speaker 3: human connections, the fact is these musicians are already under

573
00:34:11,360 --> 00:34:15,160
Speaker 3: pressure from streaming. The top one percent of artists on

574
00:34:15,239 --> 00:34:18,600
Speaker 3: Spotify and ninety percent of the royalties, meaning that for

575
00:34:18,640 --> 00:34:22,040
Speaker 3: everybody else, live music has become essential as a way

576
00:34:22,080 --> 00:34:25,320
Speaker 3: to make a musical living. But even that is getting

577
00:34:25,360 --> 00:34:29,000
Speaker 3: tough as top concert tours from superstars like Taylor Swift

578
00:34:29,360 --> 00:34:33,680
Speaker 3: get bigger, pricier, and more dominant. Singer and songwriter Taylor

579
00:34:33,800 --> 00:34:38,560
Speaker 3: Swift has a new reputation for boosting the US economy.

580
00:34:38,760 --> 00:34:41,560
Speaker 3: It's estimated that her tour could generate over four and

581
00:34:41,600 --> 00:34:47,360
Speaker 3: a half billion dollars in consumer spending. More musicians acquitting

582
00:34:47,400 --> 00:34:50,480
Speaker 3: the industry unable to make a living, and it's become

583
00:34:50,560 --> 00:34:55,200
Speaker 3: normal to see Grammy nominees working as realtors. Sam Grimley

584
00:34:55,320 --> 00:34:58,280
Speaker 3: is one musician who spent years performing with artists including

585
00:34:58,360 --> 00:35:01,520
Speaker 3: Tom Jones and Ed Sheering, but whose sins has left

586
00:35:01,560 --> 00:35:02,040
Speaker 3: the industry.

587
00:35:02,880 --> 00:35:06,480
Speaker 11: Being a touring musician is the best job in the

588
00:35:06,560 --> 00:35:09,680
Speaker 11: world for someone who's twenty five years old. But once

589
00:35:09,719 --> 00:35:12,000
Speaker 11: you get to a certain stage in life, the sort

590
00:35:12,000 --> 00:35:16,319
Speaker 11: of glamour starts to fade and you start to want

591
00:35:16,360 --> 00:35:18,280
Speaker 11: slightly different things in life. So I wanted a slightly

592
00:35:18,320 --> 00:35:23,520
Speaker 11: more stable existence. When I launched my music career, it

593
00:35:23,640 --> 00:35:26,959
Speaker 11: was around the time that Napster was turning the music

594
00:35:27,000 --> 00:35:32,680
Speaker 11: industry upside down, and over the next fifteen years roughly

595
00:35:33,320 --> 00:35:40,520
Speaker 11: the overall revenue from recorded music absolutely plummeted. So in

596
00:35:40,680 --> 00:35:43,239
Speaker 11: a sense, I was lucky that I had gone into

597
00:35:43,480 --> 00:35:46,719
Speaker 11: live music industry. But another way of looking at that

598
00:35:46,880 --> 00:35:50,239
Speaker 11: is it became increasingly the only option for musicians to

599
00:35:50,640 --> 00:35:51,640
Speaker 11: make a viable living.

600
00:35:57,840 --> 00:36:00,600
Speaker 3: So if there is an existential threat out there, it's

601
00:36:00,640 --> 00:36:05,120
Speaker 3: that the combined forces of streaming and AI will break

602
00:36:05,320 --> 00:36:09,160
Speaker 3: music's long running passion principle. This is the force that

603
00:36:09,320 --> 00:36:12,360
Speaker 3: keeps new musicians coming up and dreaming that they just

604
00:36:12,520 --> 00:36:15,000
Speaker 3: might have a shot and making it big, even if

605
00:36:15,080 --> 00:36:18,160
Speaker 3: so many of them don't. If AI music is so

606
00:36:18,280 --> 00:36:22,440
Speaker 3: successful that it's smothers platforms like Spotify and disincentivizes new

607
00:36:22,520 --> 00:36:24,680
Speaker 3: music completely, that would be a disaster.

608
00:36:25,320 --> 00:36:29,200
Speaker 11: I do think AI could replace some aspects of what

609
00:36:29,719 --> 00:36:34,480
Speaker 11: musicians currently do. It seems entirely plausible that in the

610
00:36:35,040 --> 00:36:38,200
Speaker 11: very near future we could end up with AI session

611
00:36:38,280 --> 00:36:42,160
Speaker 11: musicians who are able to replicate the feel or the

612
00:36:42,360 --> 00:36:47,120
Speaker 11: sound of various famous musicians, and that really does represent

613
00:36:47,239 --> 00:36:51,520
Speaker 11: a serious threat to people for whom session performance is

614
00:36:51,600 --> 00:36:52,480
Speaker 11: their livelihood.

615
00:36:52,880 --> 00:36:55,440
Speaker 3: The poetic twist is that Sam is now a lawyer

616
00:36:55,800 --> 00:36:58,920
Speaker 3: and he talked to us about copyright, an issue that seems,

617
00:36:59,320 --> 00:37:02,440
Speaker 3: let's face it, a little dull, but it's actually absolutely

618
00:37:02,520 --> 00:37:06,120
Speaker 3: critical to the AI story. Copyright protects the ability to

619
00:37:06,160 --> 00:37:09,040
Speaker 3: profit from art, and it also incentivizes the creation of

620
00:37:09,160 --> 00:37:12,520
Speaker 3: new works. Maybe we need to strengthen it or adapt

621
00:37:12,600 --> 00:37:15,000
Speaker 3: it to AI proof future generations.

622
00:37:15,719 --> 00:37:19,879
Speaker 11: Those AI systems are trained on human creations, so there's

623
00:37:19,880 --> 00:37:24,040
Speaker 11: a question mark over whether those creators would or could

624
00:37:24,200 --> 00:37:30,920
Speaker 11: be adequately credited, compensated, or able to control the output

625
00:37:30,960 --> 00:37:31,800
Speaker 11: of those systems.

626
00:37:32,480 --> 00:37:35,600
Speaker 3: For example, imagine an AI trained on the Rolling Stones

627
00:37:35,640 --> 00:37:39,839
Speaker 3: that outputs a piece that sounds like but isn't actually satisfaction.

628
00:37:40,840 --> 00:37:44,200
Speaker 3: Might the rolling Stones have a case for compensation. Maybe

629
00:37:44,239 --> 00:37:46,520
Speaker 3: we need new rights to those musical stems that go

630
00:37:46,600 --> 00:37:50,760
Speaker 3: into an AI or an artist's general feel or groove.

631
00:37:51,480 --> 00:37:53,160
Speaker 3: And that is where Rick Astley comes in.

632
00:37:53,560 --> 00:37:57,719
Speaker 11: There's a very interesting case being brought in California by

633
00:37:58,160 --> 00:38:02,759
Speaker 11: Rick Astley, and he's essentially claiming that imitation of his

634
00:38:03,040 --> 00:38:07,720
Speaker 11: voice is an infringement of his right. An AI system

635
00:38:08,200 --> 00:38:12,040
Speaker 11: is trained to sound like or trained to have the

636
00:38:12,160 --> 00:38:16,120
Speaker 11: feel of I don't know, Ringo Star, could Ringo Star

637
00:38:16,680 --> 00:38:21,360
Speaker 11: demand a royalty fee, or demand a license fee, or

638
00:38:21,480 --> 00:38:23,200
Speaker 11: prevent the usage of that track?

639
00:38:23,920 --> 00:38:25,880
Speaker 3: To be clear, that Rick Asley case refers to a

640
00:38:26,040 --> 00:38:30,800
Speaker 3: human imitating his voice, not AI and Ashley also settled

641
00:38:30,800 --> 00:38:33,839
Speaker 3: out of court, but the fundamental legal principle remains the same.

642
00:38:34,480 --> 00:38:37,480
Speaker 3: Record labels are actually pushing for a new federal right

643
00:38:37,560 --> 00:38:41,319
Speaker 3: of publicity law that would protect artists against the unauthorized

644
00:38:41,400 --> 00:38:44,440
Speaker 3: use of their likeness or identity. Another way to AI

645
00:38:44,560 --> 00:38:48,000
Speaker 3: proof copyright is by considering the possibility of some kind

646
00:38:48,040 --> 00:38:51,200
Speaker 3: of junior credit to the machine itself. This has been

647
00:38:51,239 --> 00:38:54,560
Speaker 3: explicitly ruled out by the recent Screenwriters Union deal in Hollywood,

648
00:38:54,800 --> 00:38:58,000
Speaker 3: but in music things aren't settled yet. Crispin Hunt, who

649
00:38:58,040 --> 00:39:00,480
Speaker 3: we heard from earlier in the show, thinks this might

650
00:39:00,560 --> 00:39:04,239
Speaker 3: help us also re evaluate the role of bots and

651
00:39:04,440 --> 00:39:08,319
Speaker 3: AI and perhaps allow us to pay their human counterparts more.

652
00:39:08,760 --> 00:39:11,520
Speaker 8: I believe we need to give AI some form of

653
00:39:12,239 --> 00:39:16,200
Speaker 8: formalized not quite copyright, but maybe a neighboring right so

654
00:39:16,360 --> 00:39:20,640
Speaker 8: that it has a value. Otherwise it will be completely free,

655
00:39:20,880 --> 00:39:24,200
Speaker 8: and that will completely undermine the possibility of humans making

656
00:39:24,239 --> 00:39:25,160
Speaker 8: a living from music.

657
00:39:26,200 --> 00:39:29,399
Speaker 3: Some of that is getting closer. The website Boomy says

658
00:39:29,440 --> 00:39:32,920
Speaker 3: that music produced using its platform belongs to Boomy, and

659
00:39:33,080 --> 00:39:36,360
Speaker 3: albums produced by Endel carry its name on the songwriting credits.

660
00:39:37,160 --> 00:39:40,279
Speaker 3: If that were also accompany by better transparency on what

661
00:39:40,480 --> 00:39:43,080
Speaker 3: is AI and what is human music, we could be

662
00:39:43,120 --> 00:39:45,839
Speaker 3: at the start of a more positive debate about better

663
00:39:46,000 --> 00:39:48,840
Speaker 3: rewarding authentic human creativity.

664
00:39:49,440 --> 00:39:52,000
Speaker 8: I think it would be fairly straightforward for the streaming

665
00:39:52,080 --> 00:39:54,080
Speaker 8: services to have a button that you can flip where

666
00:39:54,080 --> 00:39:58,480
Speaker 8: you listen to only sixty percent created human music, or

667
00:39:58,680 --> 00:40:00,799
Speaker 8: you can set a level if if the robot music

668
00:40:00,880 --> 00:40:04,680
Speaker 8: has some kind of neighboring right or copyright, then I

669
00:40:04,719 --> 00:40:07,040
Speaker 8: don't think the streaming services can charge as much. I

670
00:40:07,080 --> 00:40:09,520
Speaker 8: think they can charge twenty quid a month for organic

671
00:40:09,640 --> 00:40:12,279
Speaker 8: human made music and three quid a month if you're

672
00:40:12,320 --> 00:40:13,640
Speaker 8: going to be listening to robots.

673
00:40:14,440 --> 00:40:18,360
Speaker 3: Crispin's idea could be getting closer to reality. Universal Music's

674
00:40:18,400 --> 00:40:22,399
Speaker 3: partnership with Deezer proposes to pay human artists more above

675
00:40:22,400 --> 00:40:26,000
Speaker 3: a certain threshold, but more needs to be done. This

676
00:40:26,160 --> 00:40:28,399
Speaker 3: could be a good moment to consider a whole new

677
00:40:28,520 --> 00:40:32,120
Speaker 3: payment model for streaming, such as the user centering model,

678
00:40:32,520 --> 00:40:36,520
Speaker 3: which would take everybody's individual subscription fees and pay them

679
00:40:36,560 --> 00:40:39,800
Speaker 3: out according to what they actually listen to themselves on Spotify.

680
00:40:40,440 --> 00:40:43,200
Speaker 3: The current system puts all of the fees into one

681
00:40:43,280 --> 00:40:47,000
Speaker 3: pot and distributes them according to market share. So I

682
00:40:47,040 --> 00:40:50,080
Speaker 3: think ultimately, even as there is so much to grapple

683
00:40:50,160 --> 00:40:53,480
Speaker 3: with when it comes to AI, there's one big positive.

684
00:40:53,800 --> 00:40:56,160
Speaker 3: It just might force us to put a new value

685
00:40:56,200 --> 00:40:59,640
Speaker 3: on what's authentically human and that might just be the

686
00:40:59,719 --> 00:41:04,440
Speaker 3: third way between utopia and dystopia that keeps artists stepping

687
00:41:04,520 --> 00:41:05,080
Speaker 3: up to the plane.

688
00:41:05,800 --> 00:41:08,759
Speaker 2: Okay, Leonel, I have to jump in again. As you know,

689
00:41:09,160 --> 00:41:11,960
Speaker 2: crash course is all about learning something new, and we

690
00:41:12,040 --> 00:41:15,080
Speaker 2: don't let anybody get off the show without telling us

691
00:41:15,120 --> 00:41:17,320
Speaker 2: what they've learned. So what were your key takeaways?

692
00:41:17,800 --> 00:41:17,840
Speaker 9: So?

693
00:41:17,920 --> 00:41:20,640
Speaker 3: I guess one takeaway is that AI is already hit,

694
00:41:20,840 --> 00:41:23,400
Speaker 3: so the debate needs to shift to how to use

695
00:41:23,440 --> 00:41:27,000
Speaker 3: it safely and also how to change the industry's economics

696
00:41:27,120 --> 00:41:30,680
Speaker 3: to put human artists first. I think what isn't clear

697
00:41:30,920 --> 00:41:33,760
Speaker 3: is the actual level of demand that exists for AI music.

698
00:41:34,000 --> 00:41:36,000
Speaker 3: I mean Napster and made everyone feel like a kid

699
00:41:36,040 --> 00:41:38,320
Speaker 3: in a candy stool. This feels a little more niche

700
00:41:38,560 --> 00:41:43,000
Speaker 3: like VR or even the metaverse. I think, though, it's

701
00:41:43,080 --> 00:41:45,880
Speaker 3: clear that tech is taking over more and more of

702
00:41:46,080 --> 00:41:49,320
Speaker 3: music and culture, and that is going to be what

703
00:41:49,440 --> 00:41:51,560
Speaker 3: ultimately holds lessons for the rest of the economy.

704
00:41:52,320 --> 00:41:55,200
Speaker 2: All right, Leonelle, that's a wrap. Thanks for joining us today.

705
00:41:55,640 --> 00:41:57,000
Speaker 3: Thanks for having me. This was fun.

706
00:42:00,080 --> 00:42:04,480
Speaker 2: Crash course. We believe the collisions can be messy, impressive, challenging, surprising,

707
00:42:04,960 --> 00:42:08,680
Speaker 2: and always instructive. In today's crash Course, I learned that

708
00:42:08,719 --> 00:42:11,319
Speaker 2: while AI is turning the music industry on its head,

709
00:42:11,640 --> 00:42:14,160
Speaker 2: as it is so many other industries, there's also some

710
00:42:14,400 --> 00:42:17,560
Speaker 2: very smart and creative people trying to figure out how

711
00:42:17,560 --> 00:42:21,040
Speaker 2: to navigate around some of those problems. What did you learn?

712
00:42:21,440 --> 00:42:23,600
Speaker 2: We'd love to hear from you. You can tweet at

713
00:42:23,600 --> 00:42:27,319
Speaker 2: the Bloomberg Opinion handle at Opinion or me at Tim

714
00:42:27,360 --> 00:42:31,560
Speaker 2: O'Brien using the hashtag Bloomberg Crash Course. You can also

715
00:42:31,600 --> 00:42:34,840
Speaker 2: subscribe to our show wherever you're listening right now, and

716
00:42:34,920 --> 00:42:37,520
Speaker 2: please leave us a review. It helps more people find

717
00:42:37,560 --> 00:42:41,840
Speaker 2: the show. This episode was produced by Moses Adam, Linael

718
00:42:41,920 --> 00:42:46,840
Speaker 2: Laurant and Anna mas Rakas. Our supervising producer is Magnus Hendrickson,

719
00:42:47,280 --> 00:42:50,759
Speaker 2: and we had editing help from Sagebauman, Jeff Grocott, Mike

720
00:42:50,880 --> 00:42:54,680
Speaker 2: Nitze and Christine Vanden Bilart. Blake Maple says. Our sound

721
00:42:54,760 --> 00:42:58,319
Speaker 2: engineering and our original theme song was composed by Luis Kara.

722
00:42:59,040 --> 00:43:01,759
Speaker 2: I'm Tim O'Brien. We will be back next week with

723
00:43:01,840 --> 00:43:02,800
Speaker 2: another Crash course