1 00:00:01,320 --> 00:00:06,160 Speaker 1: And Amanda gem Nation the Daily o. Emma Gillespie is here. 2 00:00:06,400 --> 00:00:08,280 Speaker 1: AI music, Well. 3 00:00:08,160 --> 00:00:11,680 Speaker 2: I have one of the great philosophical questions of our time. 4 00:00:12,320 --> 00:00:15,520 Speaker 2: Are you listening to a real artist or a robot? 5 00:00:15,920 --> 00:00:16,239 Speaker 1: Music? 6 00:00:16,280 --> 00:00:19,280 Speaker 2: AI, of course, is a huge topic of conversation at 7 00:00:19,320 --> 00:00:22,280 Speaker 2: the moment. There's a big study that's come out, a 8 00:00:22,400 --> 00:00:26,000 Speaker 2: global survey from around the world ten thousand people who 9 00:00:26,079 --> 00:00:29,840 Speaker 2: were asked if they could correctly identify whether a track 10 00:00:30,000 --> 00:00:33,879 Speaker 2: was fully AI generated or fully human generated. So I 11 00:00:33,960 --> 00:00:38,400 Speaker 2: thought we'd play little game. I'm gonna play you two songs. 12 00:00:38,920 --> 00:00:41,400 Speaker 2: I want you to tell me which one you think 13 00:00:41,600 --> 00:00:43,960 Speaker 2: is AI. So one of them is an AI track, 14 00:00:44,240 --> 00:00:47,000 Speaker 2: one of them is a real human person track. So 15 00:00:47,040 --> 00:00:49,080 Speaker 2: you've got option A, which is this one? 16 00:00:50,560 --> 00:00:51,519 Speaker 3: If you don't like. 17 00:00:52,960 --> 00:00:56,240 Speaker 1: I'm a key walk, I like it. 18 00:00:56,360 --> 00:00:59,480 Speaker 3: Change changer man changing my song? 19 00:01:03,000 --> 00:01:05,360 Speaker 1: I like it. Yeah, it's pretty good. Right, that's a 20 00:01:05,920 --> 00:01:17,520 Speaker 1: but is it aile real? No comment? Here's b see 21 00:01:17,560 --> 00:01:19,600 Speaker 1: I happened. I went to see this man in concert, 22 00:01:20,840 --> 00:01:25,119 Speaker 1: so I know. Sorry to record for you. This one 23 00:01:25,200 --> 00:01:29,520 Speaker 1: is Chris Stapleton. Well, I was going to say that 24 00:01:29,560 --> 00:01:31,640 Speaker 1: the first one was AI. I thought the first one 25 00:01:31,720 --> 00:01:33,720 Speaker 1: was AI. And that's Chris Stableton. 26 00:01:33,840 --> 00:01:36,720 Speaker 2: Well, congratulations, you were in the top three percent of 27 00:01:36,760 --> 00:01:37,839 Speaker 2: the populace. 28 00:01:37,440 --> 00:01:40,200 Speaker 1: Probably because we knew the christ. We do this for 29 00:01:40,280 --> 00:01:41,839 Speaker 1: a job. I've got to listen. 30 00:01:42,360 --> 00:01:44,440 Speaker 3: But if you didn't ask us to be quizzed on that, 31 00:01:44,840 --> 00:01:47,600 Speaker 3: I wouldn't have known that that first song, you could easily. 32 00:01:47,360 --> 00:01:49,640 Speaker 1: Have just listened to the first ever thought anything? And 33 00:01:49,800 --> 00:01:52,080 Speaker 1: so is that who's put that together? So that is 34 00:01:52,080 --> 00:01:52,880 Speaker 1: an AI song. 35 00:01:52,960 --> 00:01:56,840 Speaker 2: It's called Walk My Walk by an artist called Breaking Rust. 36 00:01:57,320 --> 00:02:01,440 Speaker 2: It was number one on the Billboard Country Charts last week. 37 00:02:01,720 --> 00:02:06,360 Speaker 2: Really lately AI generated Oh, this artist Breaking Rast, who 38 00:02:06,400 --> 00:02:09,880 Speaker 2: doesn't exist, has over two point three million monthly listeners 39 00:02:09,880 --> 00:02:12,640 Speaker 2: on Spotify. That song that we heard that little clip 40 00:02:12,639 --> 00:02:16,280 Speaker 2: of has three point six million streams to date. 41 00:02:16,600 --> 00:02:20,440 Speaker 1: Is it a secret that he's AI. Well, that's the question. 42 00:02:20,520 --> 00:02:24,080 Speaker 2: I guess it's not obvious on Spotify anywhere. You have 43 00:02:24,120 --> 00:02:26,680 Speaker 2: to go into the credits of the song to find 44 00:02:26,720 --> 00:02:28,800 Speaker 2: the names of the songwriters and then if you look 45 00:02:28,840 --> 00:02:32,359 Speaker 2: those people up, you can quickly realize that they're not real. 46 00:02:32,720 --> 00:02:34,000 Speaker 2: But this is part of the problem. 47 00:02:34,040 --> 00:02:34,520 Speaker 1: At the moment. 48 00:02:34,560 --> 00:02:38,280 Speaker 2: There was this Reuter's Global survey of nine thousand people 49 00:02:38,320 --> 00:02:41,680 Speaker 2: from eight countries ninety seven percent couldn't correctly identify an 50 00:02:41,680 --> 00:02:45,239 Speaker 2: AI track, but seventy five percent said they want clear 51 00:02:45,440 --> 00:02:48,959 Speaker 2: labels on AI generated music so that they can see 52 00:02:49,040 --> 00:02:51,160 Speaker 2: those songs they can't identify as being AI, so they 53 00:02:51,200 --> 00:02:55,040 Speaker 2: have a quick and easy way of recognizing and acknowledging that. Okay, cool, 54 00:02:55,280 --> 00:02:57,440 Speaker 2: we like the way that sounded, but it wasn't made 55 00:02:57,520 --> 00:02:58,399 Speaker 2: by a person. 56 00:02:58,600 --> 00:02:59,480 Speaker 1: Do people care? 57 00:03:00,320 --> 00:03:04,440 Speaker 2: People do care, They care greatly. Sixty five percent think 58 00:03:04,520 --> 00:03:07,760 Speaker 2: that AI models shouldn't be able to be trained using 59 00:03:08,080 --> 00:03:12,600 Speaker 2: copyrighted music. Seventy percent believe that AI threatens the livelihood 60 00:03:12,600 --> 00:03:16,160 Speaker 2: of musicians, and over half said they feel uncomfortable not 61 00:03:16,360 --> 00:03:20,120 Speaker 2: knowing if they're listening to a machine generated piece of content. 62 00:03:20,520 --> 00:03:23,160 Speaker 2: So I guess there are these huge ethical concerns around 63 00:03:23,200 --> 00:03:27,280 Speaker 2: it that are ongoing, the legal concerns around copyrighted music. 64 00:03:27,040 --> 00:03:29,000 Speaker 1: Being used to heay to train it. 65 00:03:29,160 --> 00:03:32,840 Speaker 2: Well. Exactly Universal did a massive deal with an AI 66 00:03:32,919 --> 00:03:36,160 Speaker 2: song generator just a couple of months ago, and at 67 00:03:36,320 --> 00:03:38,680 Speaker 2: Universal Artists an'ch or what that means for them? Does 68 00:03:38,720 --> 00:03:41,120 Speaker 2: that mean that their songs are kind of being offered 69 00:03:41,200 --> 00:03:43,839 Speaker 2: up to the AI song generator gods and it's out 70 00:03:43,840 --> 00:03:46,720 Speaker 2: of their control. So there's a huge lack of trust 71 00:03:46,720 --> 00:03:49,080 Speaker 2: and transparency clearly over this issue. 72 00:03:49,160 --> 00:03:50,840 Speaker 3: So is this like say you're a musician and you 73 00:03:51,480 --> 00:03:55,160 Speaker 3: signed a universal and they take your song to train AI. Yeah, 74 00:03:55,360 --> 00:03:58,680 Speaker 3: it's like painting by numbers. It gets broken down and 75 00:03:58,720 --> 00:04:00,760 Speaker 3: then re constant ted it somewhere. 76 00:04:00,800 --> 00:04:01,040 Speaker 1: Yeah. 77 00:04:01,040 --> 00:04:03,160 Speaker 2: And if you say John Farnam, who we just heard from, 78 00:04:03,200 --> 00:04:05,560 Speaker 2: if you put into a song generator, I want a 79 00:04:05,600 --> 00:04:07,720 Speaker 2: song in the style of John Farnham with a voice 80 00:04:07,760 --> 00:04:11,080 Speaker 2: that sounds like his singing about the Sydney Harbor Bridge 81 00:04:11,160 --> 00:04:13,680 Speaker 2: or whatever. You know that that song could be generated 82 00:04:13,720 --> 00:04:17,040 Speaker 2: and sound just like him. But the BBC I love 83 00:04:17,120 --> 00:04:19,200 Speaker 2: This released a new article with a bit of a 84 00:04:19,279 --> 00:04:22,120 Speaker 2: kind of how to train yourself to detect AI music 85 00:04:22,160 --> 00:04:25,440 Speaker 2: with a few good tips. They recommend listening for high 86 00:04:25,480 --> 00:04:29,440 Speaker 2: frequency fuzziness. So some AI songs the vocals or the 87 00:04:29,520 --> 00:04:33,600 Speaker 2: synths have this unnatural kind of fuzzy sound around them 88 00:04:33,680 --> 00:04:37,240 Speaker 2: or high frequencies. They say to always check the artist's profile, 89 00:04:37,320 --> 00:04:39,480 Speaker 2: So if you're skeptical of a song, look up the 90 00:04:39,560 --> 00:04:40,680 Speaker 2: name of the artists, look at it. 91 00:04:40,720 --> 00:04:42,600 Speaker 1: Don't They have fake profiles too. 92 00:04:42,720 --> 00:04:45,480 Speaker 2: Yes, but you can tell if you look close enough 93 00:04:45,480 --> 00:04:48,120 Speaker 2: at those images. Of the AI generated artists. They look 94 00:04:48,160 --> 00:04:51,080 Speaker 2: a bit cartoon they're a little bit cartoonish. You might 95 00:04:51,120 --> 00:04:53,360 Speaker 2: not think anything if you just scrolled past it quickly, 96 00:04:53,640 --> 00:04:55,040 Speaker 2: but if you look at it for longer than five 97 00:04:55,120 --> 00:04:57,520 Speaker 2: or ten seconds, you can usually tell okay, that doesn't 98 00:04:57,560 --> 00:05:00,320 Speaker 2: look like a real person. So look up the art artist, 99 00:05:00,520 --> 00:05:04,159 Speaker 2: Google them. You can use filters on some streaming platforms 100 00:05:04,560 --> 00:05:07,360 Speaker 2: to exclude AI content, so there are calls for those 101 00:05:07,360 --> 00:05:11,039 Speaker 2: filters to be more easy to use. And honestly, this 102 00:05:11,160 --> 00:05:14,080 Speaker 2: article has encouraged people to listen to a lot of 103 00:05:14,120 --> 00:05:17,120 Speaker 2: AI music because it thinks that that will help you 104 00:05:17,200 --> 00:05:19,320 Speaker 2: notice more of those subtle patterns. 105 00:05:19,000 --> 00:05:21,240 Speaker 1: And then you can discern what's real and what's not. 106 00:05:21,400 --> 00:05:28,200 Speaker 1: It's like AI radio show, who would do that? Oh 107 00:05:28,440 --> 00:05:32,680 Speaker 1: do that? He's a naughty boy, am thank you, thank you. 108 00:05:34,560 --> 00:05:36,640 Speaker 1: It's all about authentasty man.