1 00:00:05,200 --> 00:00:05,600 Speaker 1: Jiaoda. 2 00:00:05,680 --> 00:00:08,800 Speaker 2: I'm Chelsea Daniels and this is the Front Page, a 3 00:00:08,880 --> 00:00:16,680 Speaker 2: daily podcast presented by The New Zealand Herald. Artificial intelligence 4 00:00:16,680 --> 00:00:19,440 Speaker 2: has been the hot topic of debate in the business 5 00:00:19,440 --> 00:00:23,400 Speaker 2: world for the last few years, but increasingly it's an 6 00:00:23,440 --> 00:00:27,520 Speaker 2: area that is encroaching on the creative industries. The latest 7 00:00:27,560 --> 00:00:31,440 Speaker 2: open ai update is so advanced fans online have used 8 00:00:31,440 --> 00:00:35,080 Speaker 2: it to eerily replicate the hand drawn art style of 9 00:00:35,200 --> 00:00:39,479 Speaker 2: Japanese anime favorites Studio Ghibli. It's just the latest sign 10 00:00:39,560 --> 00:00:43,040 Speaker 2: of AI coming for the arts, with recent headlines also 11 00:00:43,159 --> 00:00:48,239 Speaker 2: highlighting concerns over entirely artificial models in ad campaigns and 12 00:00:48,560 --> 00:00:51,720 Speaker 2: fake movie trailers that look close to the real thing. 13 00:00:52,280 --> 00:00:55,800 Speaker 2: What protections are there in place for our creative sector 14 00:00:56,160 --> 00:00:59,120 Speaker 2: or could they become one of the first industries to 15 00:00:59,280 --> 00:01:03,120 Speaker 2: fall to our new AI overlords. Today on the Front Page, 16 00:01:03,240 --> 00:01:07,280 Speaker 2: University of Sydney Business School Associate Professor Sandra Peter is 17 00:01:07,319 --> 00:01:09,800 Speaker 2: with us to take us through the impact of. 18 00:01:09,720 --> 00:01:11,399 Speaker 1: These emerging technologies. 19 00:01:16,800 --> 00:01:21,240 Speaker 2: Sandra, you've recently written about open aiy's latest update. Can 20 00:01:21,280 --> 00:01:25,200 Speaker 2: you explain what's happening here with these Studio Ghibli style images? 21 00:01:25,560 --> 00:01:29,640 Speaker 3: So Open Eye's latest update to chat jipt was a 22 00:01:29,760 --> 00:01:35,520 Speaker 3: significantly improved image generation capability. This means that it allowed 23 00:01:35,600 --> 00:01:40,280 Speaker 3: users to create like really convincing images in the style 24 00:01:40,360 --> 00:01:45,040 Speaker 3: off And everyone's been trying out the Ghibli style because 25 00:01:45,040 --> 00:01:49,200 Speaker 3: it's been one of those very very clearly identifiable styles, 26 00:01:49,200 --> 00:01:51,600 Speaker 3: and it was also used by some altmann to tell 27 00:01:51,640 --> 00:01:54,920 Speaker 3: people about this thing, and it's been really enormously popular, 28 00:01:55,080 --> 00:01:58,960 Speaker 3: so much so that basically their systems have crashed since 29 00:01:59,000 --> 00:02:01,480 Speaker 3: they've released it. And I think it was about another 30 00:02:02,560 --> 00:02:06,280 Speaker 3: ten million people that have joined the open aies efforts 31 00:02:06,560 --> 00:02:09,440 Speaker 3: based on this new image generation model. 32 00:02:09,600 --> 00:02:11,119 Speaker 1: And so how is this possible? 33 00:02:11,200 --> 00:02:14,800 Speaker 2: Has openaiy just scraped all two dozen Ghibli movies in 34 00:02:14,919 --> 00:02:16,800 Speaker 2: order to replicate this style. 35 00:02:17,639 --> 00:02:20,840 Speaker 3: They haven't. They've scraped the Internet, and the internet has 36 00:02:20,880 --> 00:02:23,840 Speaker 3: a lot of things on the internet, and we can 37 00:02:23,880 --> 00:02:27,440 Speaker 3: talk about what goes into training some of these models, 38 00:02:27,600 --> 00:02:31,440 Speaker 3: but the idea with Generative AI is that they've changed 39 00:02:31,440 --> 00:02:34,079 Speaker 3: the way that they create these images. So they've kind 40 00:02:34,120 --> 00:02:38,720 Speaker 3: of moved from the traditional what we call diffusion models, 41 00:02:38,760 --> 00:02:43,240 Speaker 3: where we have models that gradually refine just noisy data 42 00:02:43,320 --> 00:02:48,440 Speaker 3: into something called autoregressive algorithms. And no one wants to 43 00:02:48,480 --> 00:02:50,800 Speaker 3: go into the details of that. All you need to 44 00:02:50,840 --> 00:02:55,120 Speaker 3: know about is is that basically it treats images like language. 45 00:02:55,280 --> 00:02:59,680 Speaker 3: So Chagipiti now predicts words in a sentence, but it 46 00:02:59,720 --> 00:03:03,720 Speaker 3: can also predict visual elements in an image. This means 47 00:03:03,720 --> 00:03:07,160 Speaker 3: that you can basically use all the things that jipt 48 00:03:07,360 --> 00:03:11,520 Speaker 3: has learned about Ghibli style, for instance, the fact that 49 00:03:11,560 --> 00:03:15,959 Speaker 3: it's got things like soft pastels, it's a Japanese type animation, 50 00:03:16,120 --> 00:03:18,440 Speaker 3: and so on, and it can use those to more 51 00:03:18,480 --> 00:03:22,960 Speaker 3: accurately create these images from precise prompts that people give it. 52 00:03:23,200 --> 00:03:26,519 Speaker 2: What are the copyright implications here of this happening. 53 00:03:26,639 --> 00:03:29,160 Speaker 3: Oh, there's a lot of copyright images. And can I 54 00:03:29,160 --> 00:03:31,960 Speaker 3: actual say when it comes to generative AI, this comes 55 00:03:31,960 --> 00:03:36,000 Speaker 3: on top of many other controversies that we've had. But 56 00:03:36,280 --> 00:03:40,960 Speaker 3: the ability to work with styles, and when I say styles, 57 00:03:41,000 --> 00:03:44,040 Speaker 3: I don't just mean the Ghibli style. Everything becomes a 58 00:03:44,080 --> 00:03:49,600 Speaker 3: style for generative AI. That means things like bananas or cats, 59 00:03:49,680 --> 00:03:54,000 Speaker 3: or bloody corporate emails, they all become styles. The ability 60 00:03:54,040 --> 00:03:57,000 Speaker 3: to work with these styles it becomes the heart of 61 00:03:57,000 --> 00:04:01,920 Speaker 3: this controversy because for many artists, like our distinctive approaches 62 00:04:01,960 --> 00:04:05,200 Speaker 3: to how we create art is not a style. That 63 00:04:05,280 --> 00:04:09,280 Speaker 3: can be applied in a specific prompt. However, the traditionally 64 00:04:09,760 --> 00:04:14,160 Speaker 3: the copyright law doesn't protect styles, only very specific expressions 65 00:04:14,240 --> 00:04:17,440 Speaker 3: because we don't want to stifle creative expression. Right if 66 00:04:17,480 --> 00:04:21,080 Speaker 3: you could copyright things like impressionism, that would limit what 67 00:04:21,120 --> 00:04:24,799 Speaker 3: people can do. But there's a very clear difference between 68 00:04:24,800 --> 00:04:27,760 Speaker 3: a general style and then the highly distinctive style that 69 00:04:27,839 --> 00:04:32,000 Speaker 3: a person might have. If you remember a while back, 70 00:04:32,040 --> 00:04:35,520 Speaker 3: there was a guy called Greg Rutowski. This was a 71 00:04:35,520 --> 00:04:38,520 Speaker 3: Polish artist, and people were using his style over and 72 00:04:38,560 --> 00:04:42,160 Speaker 3: over again to generate images on unstable diffusion. Now, if 73 00:04:42,560 --> 00:04:45,839 Speaker 3: you try to do them in the style of Greg, 74 00:04:45,920 --> 00:04:51,760 Speaker 3: this threatening sposed his livelihood and his craft. So creators 75 00:04:51,800 --> 00:04:54,840 Speaker 3: have taken legal action against this on a number of fronts. 76 00:04:59,400 --> 00:05:02,400 Speaker 4: The newest trend is to create studio I say create 77 00:05:02,640 --> 00:05:06,080 Speaker 4: is to imitate Studio ghibli images truly just coming for 78 00:05:06,160 --> 00:05:10,360 Speaker 4: one of our most beloved examples of human manual creativity. 79 00:05:10,480 --> 00:05:12,920 Speaker 4: Everyone loves studio Ghibili because it's beautiful and also because 80 00:05:12,920 --> 00:05:17,000 Speaker 4: it's painstakingly created by humans who love art. And now 81 00:05:17,040 --> 00:05:20,000 Speaker 4: to actually bt is trying to imitate that and creating 82 00:05:20,040 --> 00:05:23,360 Speaker 4: images that are just worse. They're just soul less versions 83 00:05:23,400 --> 00:05:25,760 Speaker 4: of Studio Ghibli. It's also worth noting that Miyazaki, the 84 00:05:25,800 --> 00:05:28,920 Speaker 4: mastermind behind Studio Ghibli, he once said of AI, I 85 00:05:28,960 --> 00:05:31,960 Speaker 4: would never wish to incorporate this technology into my work 86 00:05:32,000 --> 00:05:34,360 Speaker 4: at all, and once when he was presented with an 87 00:05:34,360 --> 00:05:36,640 Speaker 4: example of how AI could be applied to his work, 88 00:05:36,720 --> 00:05:39,559 Speaker 4: he saw the footage and said quote, I strongly feel 89 00:05:39,600 --> 00:05:41,960 Speaker 4: that this is an insult to life itself. 90 00:05:46,080 --> 00:05:49,679 Speaker 2: It emerged a few months ago that allegedly Meta CEO 91 00:05:49,880 --> 00:05:54,200 Speaker 2: Mark Zuckerberg approved of his company using pirated versions of 92 00:05:54,320 --> 00:05:59,520 Speaker 2: copyright protected books illegally uploaded to sites like libjen to 93 00:05:59,680 --> 00:06:03,560 Speaker 2: try its AI software. If true, it seems like it 94 00:06:03,640 --> 00:06:05,919 Speaker 2: might be breaching all sorts of laws, I suppose, But 95 00:06:06,080 --> 00:06:08,560 Speaker 2: is there actually anything that can be done about it. 96 00:06:09,320 --> 00:06:11,880 Speaker 3: This is one of those huge, if true, kind of things, 97 00:06:11,880 --> 00:06:15,560 Speaker 3: and also something that is still being debated and contested. 98 00:06:15,680 --> 00:06:18,680 Speaker 3: There is a good likelihood that things like libjan were 99 00:06:18,800 --> 00:06:21,520 Speaker 3: ingested to train these models. We don't know at what 100 00:06:21,600 --> 00:06:25,800 Speaker 3: stages and at what scale. It's also true that when 101 00:06:25,839 --> 00:06:29,279 Speaker 3: we ingest what is commonly known as the Internet, we 102 00:06:29,320 --> 00:06:32,159 Speaker 3: will end up ingesting things that have copyright that we 103 00:06:32,200 --> 00:06:35,440 Speaker 3: don't necessarily think of being ingested. I think an easier 104 00:06:35,440 --> 00:06:38,719 Speaker 3: way to think about that might be the controversy around 105 00:06:38,920 --> 00:06:41,800 Speaker 3: generative AI models being able to write really really good 106 00:06:41,800 --> 00:06:46,360 Speaker 3: dialogue for movies because they've ingested OpenSubtitles dot org, which 107 00:06:46,360 --> 00:06:50,360 Speaker 3: has pirated subtitles for all the movies in all the languages, 108 00:06:50,640 --> 00:06:53,279 Speaker 3: and even though it wasn't trained on actual movies and 109 00:06:53,320 --> 00:06:57,880 Speaker 3: on movie script ingesting the pirated subtitles means that it's 110 00:06:57,880 --> 00:07:01,560 Speaker 3: now very good at doing this. Yes, things can be done. 111 00:07:02,880 --> 00:07:03,800 Speaker 1: The law will. 112 00:07:03,640 --> 00:07:07,960 Speaker 3: Evolve much as the technology has, except that technology moves 113 00:07:08,000 --> 00:07:10,320 Speaker 3: a lot faster than the law. But there's a lot 114 00:07:10,360 --> 00:07:13,600 Speaker 3: of work on the way on new legislation to try 115 00:07:13,640 --> 00:07:16,040 Speaker 3: to balance the fact that we need an enormous amount 116 00:07:16,040 --> 00:07:20,560 Speaker 3: of data to create these models. We're protecting things like artists' 117 00:07:20,680 --> 00:07:24,400 Speaker 3: identities and their creative work. And it's obviously not just 118 00:07:25,280 --> 00:07:30,400 Speaker 3: studio Ghibili right. There's same concerns in written text in music. 119 00:07:30,480 --> 00:07:33,840 Speaker 3: I've had Billie Eilish and Pearl Jam voicing concerns about 120 00:07:33,880 --> 00:07:36,840 Speaker 3: generative AI in music. So this is very widespread and 121 00:07:36,920 --> 00:07:39,880 Speaker 3: something that is being debated around the world now. The 122 00:07:39,960 --> 00:07:44,120 Speaker 3: question of how we'll be able to kind of balance 123 00:07:44,160 --> 00:07:48,040 Speaker 3: the fact that we do need a lot of data 124 00:07:48,120 --> 00:07:50,240 Speaker 3: and there is scarcity in data, the fact that this 125 00:07:50,400 --> 00:07:54,800 Speaker 3: legislation is has there are varied approaches around the globe. 126 00:07:54,880 --> 00:07:57,680 Speaker 3: There have been caused in the US to allow companies 127 00:07:58,280 --> 00:08:01,000 Speaker 3: to use copyrighted information to this go away for the 128 00:08:01,040 --> 00:08:04,920 Speaker 3: company's training generative model. That we have an AI race, 129 00:08:05,640 --> 00:08:08,680 Speaker 3: this is all this, this will all have to be 130 00:08:08,720 --> 00:08:10,360 Speaker 3: worked out in the next couple of years. 131 00:08:20,960 --> 00:08:25,960 Speaker 2: We've also recently seen the YouTube channel screen Culture demonetized. 132 00:08:26,160 --> 00:08:29,120 Speaker 2: You may have seen some of their AI generated trailers 133 00:08:29,160 --> 00:08:30,960 Speaker 2: that they make look like the real thing. 134 00:08:31,280 --> 00:08:33,559 Speaker 1: How advanced is AI getting. 135 00:08:33,720 --> 00:08:35,760 Speaker 2: Are we going to be able to tell the reels 136 00:08:35,800 --> 00:08:38,360 Speaker 2: from the fakes in a few years time, And what 137 00:08:38,400 --> 00:08:41,040 Speaker 2: does that mean for film studios and the like. 138 00:08:41,200 --> 00:08:43,959 Speaker 3: That's a that's a complex question. Let me let me, 139 00:08:44,080 --> 00:08:46,520 Speaker 3: let me, let me take a step back and talk 140 00:08:46,559 --> 00:08:50,000 Speaker 3: about the bigger picture. This is this is an arms race, 141 00:08:50,120 --> 00:08:53,240 Speaker 3: and at the moment we're not really winning it. In 142 00:08:53,280 --> 00:08:57,240 Speaker 3: that we are able to create very very good fakes, 143 00:08:57,320 --> 00:09:02,320 Speaker 3: deep fakes, and any type of content that is indistinguishable 144 00:09:03,080 --> 00:09:05,319 Speaker 3: from the real thing. We were already able to do 145 00:09:05,360 --> 00:09:08,559 Speaker 3: this really quite well with images a few years ago. 146 00:09:08,920 --> 00:09:11,640 Speaker 3: We are now increasingly good at doing it with a 147 00:09:11,960 --> 00:09:16,240 Speaker 3: voice with audio, we are increasingly good at doing it 148 00:09:16,000 --> 00:09:19,280 Speaker 3: with video. I think the thing to focus on here 149 00:09:19,559 --> 00:09:22,600 Speaker 3: is first that most of the generators, that all of 150 00:09:22,600 --> 00:09:25,800 Speaker 3: the generators that are out there online cannot reliably tell 151 00:09:25,840 --> 00:09:28,199 Speaker 3: you if a work is AI generated or not. If 152 00:09:28,240 --> 00:09:31,760 Speaker 3: you remember the Pope in a white Valenciaga jacket controversy, 153 00:09:32,280 --> 00:09:34,880 Speaker 3: it's been an arms race since, and that most of 154 00:09:34,960 --> 00:09:38,599 Speaker 3: us can create convincing content using over the counter, commercially 155 00:09:38,600 --> 00:09:41,440 Speaker 3: available software. So if you were trying to create a 156 00:09:41,679 --> 00:09:44,200 Speaker 3: deep fake of my voice, it would likely cost you 157 00:09:44,240 --> 00:09:47,160 Speaker 3: two dollars a month with commercially available software, and it 158 00:09:47,200 --> 00:09:49,839 Speaker 3: would be a fantastic clone of my voice that can 159 00:09:49,840 --> 00:09:52,600 Speaker 3: fool my mom. If you were looking to create video, 160 00:09:52,679 --> 00:09:54,800 Speaker 3: you'd probably need a minute and a half of video 161 00:09:54,840 --> 00:09:57,040 Speaker 3: of me in the public domain to create deep fakes 162 00:09:57,040 --> 00:10:00,040 Speaker 3: of me. So it is an arms race, and you 163 00:10:00,160 --> 00:10:02,320 Speaker 3: know we're not good at telling these things apart. 164 00:10:02,679 --> 00:10:04,280 Speaker 1: And it's not just movies. 165 00:10:04,679 --> 00:10:08,560 Speaker 2: Recently I saw global fashion giant H and M announced 166 00:10:08,600 --> 00:10:12,520 Speaker 2: plans to use AI by making digital twins of thirty 167 00:10:12,559 --> 00:10:14,880 Speaker 2: of its models. Now the company has said that the 168 00:10:14,920 --> 00:10:18,439 Speaker 2: models would own the rights to their likeness. But this 169 00:10:18,480 --> 00:10:22,200 Speaker 2: is something that's been worrying the fashion industry insiders for 170 00:10:22,280 --> 00:10:25,360 Speaker 2: some time. Does there need to be better regulations in 171 00:10:25,440 --> 00:10:29,040 Speaker 2: place before companies start to innovate and use AI in 172 00:10:29,080 --> 00:10:29,520 Speaker 2: this way? 173 00:10:30,120 --> 00:10:33,760 Speaker 3: Obviously there needs to be better regulation in this space, 174 00:10:33,840 --> 00:10:37,080 Speaker 3: but I think first it will be down to companies 175 00:10:37,120 --> 00:10:40,280 Speaker 3: to figure out how they want to use these technologies ethically. 176 00:10:40,400 --> 00:10:44,440 Speaker 3: Technology always evolves faster than the law, so the work 177 00:10:44,880 --> 00:10:46,960 Speaker 3: in the legal space will take some time. So I 178 00:10:46,960 --> 00:10:50,440 Speaker 3: think it's really really important for executives, for leaders in 179 00:10:50,440 --> 00:10:54,239 Speaker 3: this space to upscale themselves around artificial intelligence, to understand 180 00:10:54,320 --> 00:10:58,679 Speaker 3: what the ethical challenges are, what the practical challenges are, 181 00:10:58,679 --> 00:11:01,200 Speaker 3: but also what some of the huge opportunities are in 182 00:11:01,240 --> 00:11:03,880 Speaker 3: that space, and to make informed choices about what they 183 00:11:03,920 --> 00:11:08,040 Speaker 3: do in their organizations. This has been a long time coming, right. 184 00:11:08,080 --> 00:11:11,920 Speaker 3: We've had digital humans like like Little Mikaela back in 185 00:11:12,000 --> 00:11:15,600 Speaker 3: twenty sixteen. They were on Instagram, then they became you know, 186 00:11:15,800 --> 00:11:18,720 Speaker 3: digital flesh and blood. They did ads for things like 187 00:11:19,320 --> 00:11:21,480 Speaker 3: I think it was Prada back in the day and 188 00:11:21,960 --> 00:11:27,720 Speaker 3: bal Main. We had them evolve music careers on Spotify. 189 00:11:27,840 --> 00:11:30,480 Speaker 3: So this is not new, but it's up to companies 190 00:11:30,520 --> 00:11:33,000 Speaker 3: in the first instance, to really figure out how they 191 00:11:33,000 --> 00:11:35,040 Speaker 3: want to do this the right way, and also a 192 00:11:35,120 --> 00:11:38,000 Speaker 3: huge opportunity for companies to lead in an ethical way 193 00:11:38,000 --> 00:11:38,600 Speaker 3: in that space. 194 00:11:41,920 --> 00:11:45,680 Speaker 5: One of the most iconic brands bringing in the holiday season, 195 00:11:46,800 --> 00:11:49,320 Speaker 5: but take a closer look at the new Coca Cola 196 00:11:49,320 --> 00:11:52,320 Speaker 5: commercial and you might notice that it was made with 197 00:11:52,559 --> 00:11:57,000 Speaker 5: artificial intelligence. Social media is certainly caught with it shows 198 00:11:57,040 --> 00:12:01,120 Speaker 5: how lifeless that Christmas commercial is. Book A Colloges put 199 00:12:01,120 --> 00:12:03,720 Speaker 5: out an ad and ruin Christmas and their entire brand. 200 00:12:03,840 --> 00:12:09,040 Speaker 3: It's less festive, more creepy holiday vibes. We's push for 201 00:12:09,120 --> 00:12:12,959 Speaker 3: marketing efficiency, right, how do we create more with less? 202 00:12:13,200 --> 00:12:15,959 Speaker 1: That's just a sort of business one on one. 203 00:12:17,400 --> 00:12:22,320 Speaker 2: I've also seen, you know, entire ad campaigns generated by 204 00:12:22,440 --> 00:12:26,320 Speaker 2: using AI. Campaigns that would usually require entire teams, days 205 00:12:26,320 --> 00:12:31,199 Speaker 2: of shoots, makeup artists, stylists, lighting technicians, directors, photographers all 206 00:12:31,280 --> 00:12:34,480 Speaker 2: replaced by just typing words into this kind of generator. 207 00:12:34,520 --> 00:12:36,280 Speaker 1: There's something sad about that, isn't there. 208 00:12:36,440 --> 00:12:39,160 Speaker 3: Well, there's something sad about the volume that we can 209 00:12:39,240 --> 00:12:44,200 Speaker 3: make in this space. There again, opportunities in that space 210 00:12:44,240 --> 00:12:46,400 Speaker 3: as well. If you think about the kind of a 211 00:12:46,520 --> 00:12:49,160 Speaker 3: deep fake of me. We're not using it obviously to 212 00:12:49,200 --> 00:12:51,680 Speaker 3: create videos of me, but if we need to do 213 00:12:51,760 --> 00:12:54,560 Speaker 3: a little pick up, it might be easier to do 214 00:12:54,640 --> 00:12:57,760 Speaker 3: it with a typing word in rather than booking the 215 00:12:57,760 --> 00:13:01,000 Speaker 3: whole studio and everything else again and disrupt everybody's day. 216 00:13:01,160 --> 00:13:05,280 Speaker 3: So there are opportunities in that space, but there's also 217 00:13:05,600 --> 00:13:10,640 Speaker 3: a real danger that we might be just rehashing, recombining 218 00:13:10,679 --> 00:13:16,080 Speaker 3: all things and not giving artists, creators, directors, producers, writers 219 00:13:16,160 --> 00:13:19,400 Speaker 3: the opportunity to really bring our humanity to this. I 220 00:13:19,440 --> 00:13:22,880 Speaker 3: think the useful way to think about AI is as 221 00:13:22,960 --> 00:13:26,200 Speaker 3: an assistant. This is not a technology that should come 222 00:13:26,280 --> 00:13:29,440 Speaker 3: to replace what we do. It's not there to take 223 00:13:29,480 --> 00:13:32,440 Speaker 3: our jobs, but it's there to enhance us. So I 224 00:13:32,440 --> 00:13:34,640 Speaker 3: would encourage people to think about what they can do 225 00:13:34,720 --> 00:13:37,560 Speaker 3: with AI that we weren't able to do before, and 226 00:13:37,640 --> 00:13:40,720 Speaker 3: think of it as adding to our capabilities rather than 227 00:13:41,120 --> 00:13:44,320 Speaker 3: replacing us. There will be disruption, and I think all 228 00:13:44,400 --> 00:13:46,920 Speaker 3: of us, all of us know that that will happen. 229 00:13:47,600 --> 00:13:49,520 Speaker 3: I always say it's not coming for your job, but 230 00:13:49,600 --> 00:13:52,839 Speaker 3: it's definitely coming for your job description. But it's an 231 00:13:52,840 --> 00:13:58,880 Speaker 3: important moment, I think for especially people who lead businesses 232 00:13:59,000 --> 00:14:03,360 Speaker 3: lead creative things to upscale themselves on what this technology is. 233 00:14:04,200 --> 00:14:06,280 Speaker 3: I would, of course say the University of Sydney had 234 00:14:06,280 --> 00:14:09,160 Speaker 3: does the best work around the effluency anywhere in the world. 235 00:14:09,200 --> 00:14:12,120 Speaker 3: So do come and do this with us, but find 236 00:14:12,360 --> 00:14:15,160 Speaker 3: find ways to understand what this technology can do. It's 237 00:14:15,240 --> 00:14:18,160 Speaker 3: really quite different to what we've been able to do before. 238 00:14:18,800 --> 00:14:22,680 Speaker 3: Even the idea of style engines. It's not intuitive to 239 00:14:22,760 --> 00:14:25,360 Speaker 3: us to think that, you know, things like bananas or 240 00:14:25,400 --> 00:14:29,280 Speaker 3: corporate emos become styles. Catness is a style, right, So 241 00:14:29,840 --> 00:14:32,200 Speaker 3: try to understand what the tech can do and then 242 00:14:32,320 --> 00:14:35,840 Speaker 3: really be very mindful and very deliberate in how you 243 00:14:35,880 --> 00:14:38,520 Speaker 3: implement it in your organization. I think it's up to 244 00:14:39,160 --> 00:14:42,040 Speaker 3: people who lead, whether it's the creative industries or whether 245 00:14:42,040 --> 00:14:44,240 Speaker 3: it's any business or government, to lead the way on 246 00:14:44,360 --> 00:14:47,320 Speaker 3: how we think we ethically want to be doing this. 247 00:14:47,520 --> 00:14:50,800 Speaker 3: It's not happening to us. We are making this future 248 00:14:50,840 --> 00:14:53,720 Speaker 3: happen and the next two years will be crucial. 249 00:14:53,960 --> 00:14:56,480 Speaker 2: Are there any protections that you'd like to see in 250 00:14:56,600 --> 00:15:00,280 Speaker 2: place in order to better protect the creative industries? Because 251 00:15:00,320 --> 00:15:05,160 Speaker 2: I think that there is, given how fast the technology 252 00:15:05,320 --> 00:15:08,880 Speaker 2: is evolving, there is a fear that it could stifle 253 00:15:09,000 --> 00:15:12,240 Speaker 2: that creative production and that creative process. 254 00:15:12,520 --> 00:15:15,160 Speaker 3: I'm not a lawyer, so I live the I leave 255 00:15:15,240 --> 00:15:19,680 Speaker 3: the lawmaking and the details of this to those who 256 00:15:19,800 --> 00:15:23,440 Speaker 3: who know better. But I would want to see protections 257 00:15:23,560 --> 00:15:27,400 Speaker 3: around artists, right. I do want to see protections around 258 00:15:27,520 --> 00:15:31,640 Speaker 3: how their work is being used to train these models, 259 00:15:31,680 --> 00:15:33,680 Speaker 3: because I think at the moment it's a bit of 260 00:15:33,680 --> 00:15:38,120 Speaker 3: the wild West out there, and I am worried that 261 00:15:38,240 --> 00:15:41,600 Speaker 3: AI generated content might end up in the long term 262 00:15:41,680 --> 00:15:46,760 Speaker 3: severely diluting the earnings for original creators, which also means 263 00:15:46,760 --> 00:15:49,440 Speaker 3: that fewer people will want to join the creative art 264 00:15:49,680 --> 00:15:54,120 Speaker 3: So ultimately, I think owners consent in using any of 265 00:15:54,160 --> 00:15:56,040 Speaker 3: this should be should be a requirement. 266 00:15:56,160 --> 00:15:58,400 Speaker 1: Thanks for joining us, Sandra, Thanks for having me. 267 00:16:02,240 --> 00:16:05,320 Speaker 2: That's it for this episode of The Front Page. You 268 00:16:05,360 --> 00:16:09,160 Speaker 2: can read more about today's stories and extensive news coverage 269 00:16:09,200 --> 00:16:13,240 Speaker 2: at enziherld dot co dot nz. The Front Page is 270 00:16:13,280 --> 00:16:17,000 Speaker 2: produced by Ethan Sills and Richard Martin, who is also 271 00:16:17,160 --> 00:16:18,280 Speaker 2: a sound engineer. 272 00:16:18,720 --> 00:16:20,240 Speaker 1: I'm Chelsea Daniels. 273 00:16:20,840 --> 00:16:23,960 Speaker 2: Subscribe to the Front Page on iHeartRadio or wherever you 274 00:16:24,000 --> 00:16:27,880 Speaker 2: get your podcasts, and tune in on Monday for another 275 00:16:27,960 --> 00:16:29,600 Speaker 2: look behind the headlines.