1 00:00:01,240 --> 00:00:03,560 Speaker 1: My name is Lily Maddon and I'm a proud Arunda 2 00:00:03,760 --> 00:00:08,560 Speaker 1: Bunjelung Calcuttin woman from Gadighal Country. The Daily oz acknowledges 3 00:00:08,640 --> 00:00:10,799 Speaker 1: that this podcast is recorded on the lands of the 4 00:00:10,840 --> 00:00:14,400 Speaker 1: Gadighl people and pays respect to all Aboriginal and Torres 5 00:00:14,400 --> 00:00:17,319 Speaker 1: Straight Island and nations. We pay our respects to the 6 00:00:17,320 --> 00:00:20,120 Speaker 1: first peoples of these countries, both past and present. 7 00:00:25,720 --> 00:00:28,360 Speaker 2: Good morning and welcome to the Daily os. It's Wednesday, 8 00:00:28,360 --> 00:00:29,479 Speaker 2: the twenty first of June. 9 00:00:29,600 --> 00:00:32,000 Speaker 3: I'm zara, I'm sam. 10 00:00:31,480 --> 00:00:34,760 Speaker 2: AI and its impact on the music industry has been 11 00:00:34,760 --> 00:00:37,279 Speaker 2: in the headlines over the last few weeks. There are 12 00:00:37,360 --> 00:00:40,040 Speaker 2: new rules at the Grammys with the brave new world 13 00:00:40,080 --> 00:00:42,000 Speaker 2: of artificial intelligence and mind. 14 00:00:42,080 --> 00:00:46,160 Speaker 4: The Beatles are back, Paul Ringo, George and even John. 15 00:00:46,360 --> 00:00:48,920 Speaker 2: With the help of artificial intelligence. 16 00:00:48,320 --> 00:00:51,320 Speaker 5: We're able to take John's voice and get it pure 17 00:00:51,880 --> 00:00:52,959 Speaker 5: through this AI. 18 00:00:53,760 --> 00:00:56,520 Speaker 2: So what does this new technology mean for artists and 19 00:00:56,560 --> 00:00:59,720 Speaker 2: their work and how is the industry responding. We're going 20 00:00:59,760 --> 00:01:01,640 Speaker 2: to take make a look in today's deep dive. But 21 00:01:01,760 --> 00:01:03,800 Speaker 2: first Sam, what is making headlines. 22 00:01:05,959 --> 00:01:10,080 Speaker 3: Tasmanian Senator Jackie Lamby has referred allegations of war crimes 23 00:01:10,120 --> 00:01:13,600 Speaker 3: by senior Australian Defense Force personnel in Afghanistan to the 24 00:01:13,640 --> 00:01:17,240 Speaker 3: International Criminal Court. Lambie accused the federal government of hoping 25 00:01:17,319 --> 00:01:20,400 Speaker 3: the alleged actions would quote just go away, saying that 26 00:01:20,440 --> 00:01:22,120 Speaker 3: senior leaders got a free pass. 27 00:01:24,560 --> 00:01:27,600 Speaker 2: Federal politicians have been given a deadline of the seventeenth 28 00:01:27,680 --> 00:01:31,039 Speaker 2: of July to submit ridden arguments for and against the 29 00:01:31,080 --> 00:01:35,039 Speaker 2: Indigenous Voiced Parliament. This will be part of legally required 30 00:01:35,080 --> 00:01:39,520 Speaker 2: pamphlets distributed by the AEC, the Australian Electoral Commission ahead 31 00:01:39,520 --> 00:01:42,440 Speaker 2: of the Voice referendum, which will take place later this year. 32 00:01:44,440 --> 00:01:46,800 Speaker 3: Four people have been injured in a number of axe 33 00:01:46,800 --> 00:01:50,600 Speaker 3: attacks carried out at multiple Chinese restaurants across Auckland in 34 00:01:50,640 --> 00:01:53,760 Speaker 3: New Zealand. One person was arrested at the scene. Police 35 00:01:53,760 --> 00:01:56,960 Speaker 3: said the attack didn't appear to be racially motivated. 36 00:01:58,600 --> 00:02:01,640 Speaker 2: And the good news Five of the six authors shortlisted 37 00:02:01,680 --> 00:02:05,960 Speaker 2: for the prestigious Miles Franklin Literary Award our first time nominees. 38 00:02:06,160 --> 00:02:09,200 Speaker 2: The Australian Book Prize hands out a sixty thousand dollars 39 00:02:09,200 --> 00:02:11,320 Speaker 2: prize to the winner, who will be named next month. 40 00:02:15,240 --> 00:02:15,720 Speaker 3: What is that? 41 00:02:16,720 --> 00:02:18,360 Speaker 2: It's our new AI theme song. 42 00:02:18,800 --> 00:02:22,119 Speaker 3: It's very it's very kind of nineties broadcast news. 43 00:02:22,440 --> 00:02:24,440 Speaker 2: Yeah, it's not what you want to hear at six 44 00:02:24,480 --> 00:02:26,959 Speaker 2: am on a Monday morning. I don't think that synth 45 00:02:27,080 --> 00:02:29,280 Speaker 2: is like especially hard to listen to it. 46 00:02:29,280 --> 00:02:31,920 Speaker 3: It's if Van Alen and the newsroom got together and 47 00:02:32,120 --> 00:02:34,480 Speaker 3: had a baby. So that was made totally by aol. 48 00:02:34,520 --> 00:02:37,080 Speaker 2: It was created on a tool called music gen which 49 00:02:37,280 --> 00:02:41,120 Speaker 2: was launched by Meta so Facebook. You can prompt it 50 00:02:41,200 --> 00:02:43,959 Speaker 2: with the style of music you want, things like blues, guitar, 51 00:02:44,200 --> 00:02:47,680 Speaker 2: cinematic vibe, heavy drums, whatever you want, and then it 52 00:02:47,720 --> 00:02:51,520 Speaker 2: basically takes all of that and creates something unique in response. 53 00:02:51,639 --> 00:02:53,400 Speaker 2: So what you just heard there was what happened when 54 00:02:53,400 --> 00:02:55,520 Speaker 2: I prompted it to make a theme song for a 55 00:02:55,600 --> 00:02:56,480 Speaker 2: news podcast. 56 00:02:56,680 --> 00:02:58,440 Speaker 3: I don't think we're going to be adopting that one 57 00:02:58,520 --> 00:03:01,120 Speaker 3: anytime soon. I like our current theme. 58 00:03:00,960 --> 00:03:02,520 Speaker 2: And not because you're best friend made it. 59 00:03:02,680 --> 00:03:06,160 Speaker 3: No, exactly, not credit to Sam Waste Music. It didn't 60 00:03:06,160 --> 00:03:08,840 Speaker 3: ail our style, but these are very early days for AI. 61 00:03:09,360 --> 00:03:11,840 Speaker 2: Yeah, I think it is important to note that it 62 00:03:11,919 --> 00:03:15,240 Speaker 2: did create an original piece of music in mere minutes, 63 00:03:15,320 --> 00:03:18,800 Speaker 2: and no matter how amazing a composer is, it's impossible 64 00:03:18,840 --> 00:03:22,120 Speaker 2: to create something that quickly. Music Jen, the tool that 65 00:03:22,160 --> 00:03:25,480 Speaker 2: we used is open source, and by open source, I 66 00:03:25,560 --> 00:03:28,280 Speaker 2: mean that it's free and available for people to use 67 00:03:28,320 --> 00:03:31,440 Speaker 2: and modify, and it's out there alongside a range of 68 00:03:31,480 --> 00:03:34,600 Speaker 2: generative music tools like Boomy and voice mod. The give 69 00:03:34,639 --> 00:03:37,840 Speaker 2: amateur creators people like you and me, the ability to 70 00:03:37,920 --> 00:03:41,680 Speaker 2: create music. So the question I want to explore today then, 71 00:03:42,040 --> 00:03:45,240 Speaker 2: is how is AI changing how we actually create and 72 00:03:45,320 --> 00:03:46,120 Speaker 2: listen to music. 73 00:03:46,360 --> 00:03:50,240 Speaker 3: Whatever sector you're looking at AI, there's such unique impacts, 74 00:03:50,360 --> 00:03:54,680 Speaker 3: whether it's architecture, medicine, and news news. I think music 75 00:03:54,760 --> 00:03:57,000 Speaker 3: is a really interesting one. But before we dive into it, 76 00:03:57,200 --> 00:03:59,360 Speaker 3: what are we talking about generally when we're talking about 77 00:03:59,400 --> 00:04:00,320 Speaker 3: AI and music. 78 00:04:00,640 --> 00:04:04,000 Speaker 2: So most AI tools learn from pre existing data sets 79 00:04:04,040 --> 00:04:07,160 Speaker 2: that they're trained with. We're talking about things like sophisticated 80 00:04:07,200 --> 00:04:10,520 Speaker 2: algorithms which control the Internet and analyze a heap of 81 00:04:10,600 --> 00:04:13,760 Speaker 2: music tracks. For example, Meta has said that they trained 82 00:04:13,840 --> 00:04:17,080 Speaker 2: music jen, that thing that created the horrific song at 83 00:04:17,120 --> 00:04:21,159 Speaker 2: the beginning there using twenty thousand hours of licensed music 84 00:04:21,320 --> 00:04:25,000 Speaker 2: as well as almost four hundred thousand instrumental stock music tracks. 85 00:04:25,680 --> 00:04:28,240 Speaker 2: As it's being fed all of this data, the AI 86 00:04:28,400 --> 00:04:33,320 Speaker 2: identifies patterns across different music genres, melodies, instruments and structures, 87 00:04:33,880 --> 00:04:36,640 Speaker 2: and then it learns how to actually make that music, 88 00:04:36,680 --> 00:04:39,640 Speaker 2: how to create it. So you might have heard examples 89 00:04:39,640 --> 00:04:43,600 Speaker 2: of this with people using AI music tools to reproduce artists' 90 00:04:43,640 --> 00:04:46,360 Speaker 2: voices like Kanye or Taylor Swift. 91 00:04:46,000 --> 00:04:47,840 Speaker 3: And I'm going to let you finish here. But there 92 00:04:47,920 --> 00:04:50,840 Speaker 3: was also a track released by an AI version of 93 00:04:50,920 --> 00:04:53,640 Speaker 3: Drake and The Weekend that actually got pulled down a 94 00:04:53,640 --> 00:04:54,440 Speaker 3: couple of months ago. 95 00:04:54,720 --> 00:04:57,159 Speaker 2: Yeah, it was insane. It was called a Heart on 96 00:04:57,200 --> 00:05:00,160 Speaker 2: My Sleeve and it was made by a TikToker and 97 00:05:00,279 --> 00:05:03,640 Speaker 2: it used AI that had basically mastered how to sound 98 00:05:03,680 --> 00:05:06,599 Speaker 2: like Drake and The Weekend so that it could just 99 00:05:06,760 --> 00:05:10,839 Speaker 2: imitate both of them. It was eventually pulled because Universal Music, 100 00:05:10,839 --> 00:05:13,640 Speaker 2: who's the label that represents Drake in the Weekend, argued 101 00:05:13,680 --> 00:05:16,960 Speaker 2: that it was a breach of copyright. I mean, copyright 102 00:05:17,040 --> 00:05:19,279 Speaker 2: is just one issue and a whole slew of issues 103 00:05:19,279 --> 00:05:22,560 Speaker 2: that AI poses. There's also the question of consent for 104 00:05:22,640 --> 00:05:24,680 Speaker 2: all the music that's being used to train a lot 105 00:05:24,720 --> 00:05:25,520 Speaker 2: of these systems. 106 00:05:25,680 --> 00:05:29,200 Speaker 3: I can imagine developments like this are really destabilizing and 107 00:05:29,360 --> 00:05:32,760 Speaker 3: quite scary for artists and composers. If we take our 108 00:05:32,800 --> 00:05:35,120 Speaker 3: theme song that was modified at the top of the episode. 109 00:05:35,279 --> 00:05:37,479 Speaker 3: It was pretty intense, and again I don't think we'd 110 00:05:37,600 --> 00:05:39,920 Speaker 3: use it, but if you did get the sound right, 111 00:05:40,080 --> 00:05:42,160 Speaker 3: it would save a lot of money in time and 112 00:05:42,480 --> 00:05:45,720 Speaker 3: resources and access to composers and could really impact the 113 00:05:45,760 --> 00:05:48,119 Speaker 3: business that exists around original music. 114 00:05:48,680 --> 00:05:48,920 Speaker 4: Yeah. 115 00:05:48,960 --> 00:05:52,240 Speaker 2: I mean, there are so many implications for artists who 116 00:05:52,400 --> 00:05:56,600 Speaker 2: rely on income for composing production music. One of those 117 00:05:56,760 --> 00:06:00,279 Speaker 2: is Hayat Salim, who's a songwriter and composer. 118 00:06:00,440 --> 00:06:04,039 Speaker 5: A lot of composers live off production music. They spend 119 00:06:04,040 --> 00:06:06,680 Speaker 5: their time working on let's say an album of trader music, 120 00:06:06,800 --> 00:06:09,800 Speaker 5: ambient music. Their music is put out there and if 121 00:06:09,800 --> 00:06:13,680 Speaker 5: somebody licensed their music into a project, they get the royalties. 122 00:06:13,720 --> 00:06:17,520 Speaker 5: And some people really depend on this solely for their income. 123 00:06:18,160 --> 00:06:20,120 Speaker 5: And I think this is one of the first things 124 00:06:20,240 --> 00:06:23,760 Speaker 5: that could potentially be replaced by artificial intelligence. 125 00:06:24,200 --> 00:06:26,919 Speaker 2: And it's not just on the composer side. It also 126 00:06:27,000 --> 00:06:30,880 Speaker 2: raises content for record labels and for publishers. I'm going 127 00:06:30,880 --> 00:06:33,279 Speaker 2: to play you a clip of Jonathan Carter, who's the 128 00:06:33,279 --> 00:06:36,720 Speaker 2: head of Legal and Corporate Services at APRA AMKOS, which 129 00:06:36,760 --> 00:06:39,960 Speaker 2: is the body that looks after music licensing and royalties. 130 00:06:39,960 --> 00:06:42,640 Speaker 2: Here in Australia so that artists can get paid. 131 00:06:42,960 --> 00:06:46,800 Speaker 4: When now in an environment where the algorithms are making 132 00:06:46,960 --> 00:06:53,279 Speaker 4: very personalized recommendations, if you rely on music being pushed 133 00:06:53,320 --> 00:06:55,640 Speaker 4: to you in that way, it does take some of 134 00:06:55,680 --> 00:07:01,080 Speaker 4: the exploration and potential for discovering amazing new music. The 135 00:07:01,240 --> 00:07:04,640 Speaker 4: other concern is that a lot of those playlists and 136 00:07:04,720 --> 00:07:10,280 Speaker 4: algorithms are generated overseas, and the amount of Australian or 137 00:07:10,360 --> 00:07:15,240 Speaker 4: New Zealand local music the feature is very low. There 138 00:07:15,280 --> 00:07:18,920 Speaker 4: is no doubt that the music industry is watching it carefully. 139 00:07:19,120 --> 00:07:25,720 Speaker 4: Technology has always disrupted and enhanced the creation and business 140 00:07:25,720 --> 00:07:30,480 Speaker 4: of music, but it does require careful policy making around it. 141 00:07:31,560 --> 00:07:34,640 Speaker 3: Zara, It's really easy to see the challenges with AI, 142 00:07:34,880 --> 00:07:37,240 Speaker 3: especially when we're in this period where the law and 143 00:07:37,280 --> 00:07:40,880 Speaker 3: regulations haven't caught up to the technology and it's a 144 00:07:40,920 --> 00:07:42,840 Speaker 3: bit of the wild West out there in terms of 145 00:07:42,920 --> 00:07:46,040 Speaker 3: who's doing what and when. But this isn't the first 146 00:07:46,040 --> 00:07:48,680 Speaker 3: time that the music industry has had a shake up 147 00:07:48,760 --> 00:07:52,240 Speaker 3: because of technology. I mean, we don't even bat an 148 00:07:52,240 --> 00:07:54,160 Speaker 3: eyelid now when we see someone on a drum machine, 149 00:07:54,240 --> 00:07:57,400 Speaker 3: or we're listening to music that was simulated by a computer. 150 00:07:57,520 --> 00:07:59,440 Speaker 3: You know, there's not a violin in your MacBook, but 151 00:07:59,480 --> 00:08:02,800 Speaker 3: you can play. What are the positives here for music? 152 00:08:03,200 --> 00:08:03,360 Speaker 4: Yeah? 153 00:08:03,400 --> 00:08:05,560 Speaker 2: I mean I think that it is interesting to talk 154 00:08:05,560 --> 00:08:08,760 Speaker 2: about this because I think more generally with AI, there 155 00:08:08,760 --> 00:08:10,520 Speaker 2: has been a lot of focus on all of the 156 00:08:10,560 --> 00:08:13,400 Speaker 2: negatives and you know how to regulate it and what 157 00:08:13,440 --> 00:08:16,280 Speaker 2: comes next. I don't think we often talk about the 158 00:08:16,320 --> 00:08:20,120 Speaker 2: opportunities that it can afford. So it's been interesting to 159 00:08:20,200 --> 00:08:24,000 Speaker 2: see those musicians who are trying to incorporate AI as 160 00:08:24,080 --> 00:08:28,720 Speaker 2: part of their creative processes. One example is Grimes, who 161 00:08:28,840 --> 00:08:33,280 Speaker 2: created an AI platform called elf Tech, where she's allowing 162 00:08:33,280 --> 00:08:36,280 Speaker 2: people to create new music using her voice as long 163 00:08:36,320 --> 00:08:38,360 Speaker 2: as the people using it give her a cut of 164 00:08:38,400 --> 00:08:41,840 Speaker 2: the earnings. So it sounds like Grimes is singing your song, 165 00:08:41,960 --> 00:08:42,760 Speaker 2: but she's not. 166 00:08:43,120 --> 00:08:45,480 Speaker 3: Wow, that's a way to roll with the punches. Give 167 00:08:45,520 --> 00:08:46,840 Speaker 3: me a sense of what that sounds like. 168 00:08:46,960 --> 00:08:56,880 Speaker 2: Okay, size This one's called Cold Touch and it's made 169 00:08:56,920 --> 00:09:00,679 Speaker 2: with Grimes's AI voice by a composer called Keto, and 170 00:09:00,760 --> 00:09:03,760 Speaker 2: Grimes is into it. Speaking to The New York Times, 171 00:09:03,800 --> 00:09:05,960 Speaker 2: she said that she thought the chorus was really good 172 00:09:06,080 --> 00:09:09,040 Speaker 2: and that she could be convinced that she herself had 173 00:09:09,080 --> 00:09:09,920 Speaker 2: actually worked on it. 174 00:09:10,640 --> 00:09:13,800 Speaker 3: So clearly the industry is changing. Whether that's for better 175 00:09:13,880 --> 00:09:15,679 Speaker 3: or for worse is not so easy for us to 176 00:09:15,760 --> 00:09:20,360 Speaker 3: ascertain right now, though, How are the traditional institutions responding 177 00:09:20,400 --> 00:09:21,200 Speaker 3: to these changes? 178 00:09:21,640 --> 00:09:25,200 Speaker 2: Okay, so I'll give you a recent example. Just last week, 179 00:09:25,360 --> 00:09:29,200 Speaker 2: the Grammys announced new rules related to songs with AI. 180 00:09:29,800 --> 00:09:32,800 Speaker 2: For the first time, they've officially stated that music made 181 00:09:32,840 --> 00:09:35,840 Speaker 2: with AI can be eligible for an award, but the 182 00:09:35,880 --> 00:09:38,880 Speaker 2: disclaimer is that the song or performance must have significant 183 00:09:38,880 --> 00:09:42,319 Speaker 2: input by a human. Harvey Mason Jr. Who's the CEO 184 00:09:42,480 --> 00:09:45,400 Speaker 2: of the group behind the Grammys, said that it's important 185 00:09:45,400 --> 00:09:49,160 Speaker 2: because AI is going to absolutely, unequivocally have a hand 186 00:09:49,240 --> 00:09:52,200 Speaker 2: in shaping the future of our industry. The idea of 187 00:09:52,200 --> 00:09:54,319 Speaker 2: being caught off guard by it and not addressing it 188 00:09:54,400 --> 00:09:55,319 Speaker 2: is unacceptable. 189 00:09:55,640 --> 00:09:58,560 Speaker 3: It's going to be really interesting to see how our 190 00:09:58,760 --> 00:10:00,920 Speaker 3: consumer attitudes change to this. 191 00:10:01,040 --> 00:10:03,120 Speaker 2: Well, I know that. I mean, you've been really excited 192 00:10:03,200 --> 00:10:05,199 Speaker 2: by it. You've paved me a song in the car 193 00:10:05,640 --> 00:10:09,280 Speaker 2: that was entirely made by AI, and it sounded like 194 00:10:09,320 --> 00:10:12,400 Speaker 2: something that you would listen to and pay for ordinarily. 195 00:10:12,440 --> 00:10:14,960 Speaker 3: So this guy put in a request to an AI 196 00:10:15,040 --> 00:10:18,359 Speaker 3: generator to make a song that sounds like imagine Dragons, Coldplay, 197 00:10:18,400 --> 00:10:23,600 Speaker 3: The Luminiers, and One Republic. It is remarkably listenable. Here's 198 00:10:23,800 --> 00:10:25,240 Speaker 3: a bit of that to see us out today. 199 00:10:27,200 --> 00:10:28,880 Speaker 1: We've reached the time of the night. 200 00:10:28,960 --> 00:10:33,360 Speaker 3: It's the finals day, when the lights hit the sky 201 00:10:33,520 --> 00:10:36,680 Speaker 3: and the shimmer in your ard, your shadow on your beans. 202 00:10:37,480 --> 00:10:39,440 Speaker 2: Do ca me get a second, just a casher. 203 00:10:39,720 --> 00:10:41,960 Speaker 3: Thanks for joining us on the Daily Odds today. If 204 00:10:42,000 --> 00:10:45,280 Speaker 3: you learn something from today's episode, don't forget to hit subscribe. 205 00:10:45,440 --> 00:10:48,360 Speaker 3: So there's a TDA episode waiting for you every morning. 206 00:10:48,400 --> 00:10:50,800 Speaker 3: At the moment it is still written by human beings. 207 00:10:51,200 --> 00:10:53,800 Speaker 3: We'll be back again tomorrow. Until then, have a great day, 208 00:10:55,559 --> 00:10:57,080 Speaker 3: right get you're. 209 00:11:00,040 --> 00:11:02,640 Speaker 4: Added to the stars your team because shoot them down. 210 00:11:02,760 --> 00:11:05,440 Speaker 4: And now with them sitting round, it's all your magic. 211 00:11:05,520 --> 00:11:08,120 Speaker 4: I'm a round, catching your mouth leading 212 00:11:13,160 --> 00:11:13,480 Speaker 3: M hm.