1 00:00:00,160 --> 00:00:01,639 Speaker 1: Do you think we are in VO three? 2 00:00:02,840 --> 00:00:11,119 Speaker 2: If you cannot tell, does it matter? Thank god for playback. 3 00:00:11,160 --> 00:00:17,599 Speaker 2: I have no idea what I'm doing. How does Tom 4 00:00:17,680 --> 00:00:18,319 Speaker 2: Cruise do this? 5 00:00:19,680 --> 00:00:25,479 Speaker 1: We've got a lot of discuss here I go to 6 00:00:25,520 --> 00:00:26,279 Speaker 1: work against. 7 00:00:28,040 --> 00:00:29,160 Speaker 2: Thank god we aren't real. 8 00:00:36,560 --> 00:00:40,080 Speaker 3: Oh yes, the era of AI generated movie making is here, 9 00:00:40,120 --> 00:00:43,400 Speaker 3: and while it still sounds a bit awkward, we're just 10 00:00:43,440 --> 00:00:44,239 Speaker 3: getting started. 11 00:00:44,320 --> 00:00:46,360 Speaker 2: The clips you just heard were. 12 00:00:46,159 --> 00:00:51,160 Speaker 3: Generated by VO three and AI Video Generator, released last 13 00:00:51,159 --> 00:00:55,880 Speaker 3: week by Google. Like open ais Sora, it can generate 14 00:00:56,000 --> 00:00:59,760 Speaker 3: realistic looking video clips from text prompts, but it's claimed 15 00:00:59,760 --> 00:01:03,720 Speaker 3: to fit is that it also incorporates audio, including sound effects, 16 00:01:04,200 --> 00:01:08,279 Speaker 3: dialogue between characters, and even animal sounds, all lip synced 17 00:01:08,480 --> 00:01:12,839 Speaker 3: to the video. It became available to US Google vertexs 18 00:01:13,040 --> 00:01:16,720 Speaker 3: AI users on the AI Ultra subscription last week at 19 00:01:16,760 --> 00:01:19,920 Speaker 3: a cost of two hundred and fifty dollars US a month. 20 00:01:19,959 --> 00:01:22,840 Speaker 3: That's quite a lot for an amateur user, but for 21 00:01:23,280 --> 00:01:25,360 Speaker 3: pros that's actually quite affordable. 22 00:01:25,920 --> 00:01:27,279 Speaker 2: You can see where this is going. 23 00:01:27,720 --> 00:01:32,199 Speaker 3: AI generated adverts, short films, and, with enough computing power, 24 00:01:32,319 --> 00:01:36,200 Speaker 3: eventually feature films as well. It's just one way that 25 00:01:36,280 --> 00:01:40,800 Speaker 3: AI is set to disrupt the creative industries. This week 26 00:01:40,840 --> 00:01:43,520 Speaker 3: on the Business of Tech powered by two Degrees Business, 27 00:01:43,560 --> 00:01:47,520 Speaker 3: our one hundredth episode of the podcast, we're looking at 28 00:01:47,520 --> 00:01:50,280 Speaker 3: the sectors on the frontline of the AI revolution, from 29 00:01:50,360 --> 00:01:55,440 Speaker 3: journalism to photography, graphic design to filmmaking. I'm your host, 30 00:01:55,440 --> 00:01:57,880 Speaker 3: Peter Griffin, and on the Business of Tech, we dive 31 00:01:57,960 --> 00:02:01,800 Speaker 3: into the big ideas shaping digital future. My guest this 32 00:02:01,880 --> 00:02:07,000 Speaker 3: week is Lou Compagnoni, Director of Artificial Intelligence at Datacom, 33 00:02:07,520 --> 00:02:11,639 Speaker 3: the country's largest IT company with a significant trans Tasman 34 00:02:11,680 --> 00:02:15,520 Speaker 3: presence as well. Lou's based in Melbourne and she just 35 00:02:15,600 --> 00:02:19,560 Speaker 3: hosted two full day conferences, one in Auckland one in 36 00:02:19,639 --> 00:02:23,639 Speaker 3: Wellington on how the creative sector is embracing and sometimes 37 00:02:23,880 --> 00:02:27,840 Speaker 3: wrestling with artificial intelligence. I was at the Wellington events. 38 00:02:27,919 --> 00:02:33,680 Speaker 3: It was fascinating to hear from and talk to game developers, filmmakers, poets, 39 00:02:33,760 --> 00:02:38,480 Speaker 3: photographers and news media executives about how they're all grappling 40 00:02:38,960 --> 00:02:41,880 Speaker 3: with the rise of generative AI, which is giving people 41 00:02:42,120 --> 00:02:44,200 Speaker 3: unprecedented abilities. 42 00:02:43,639 --> 00:02:44,960 Speaker 2: To create content. 43 00:02:45,800 --> 00:02:50,720 Speaker 3: Lou, who joined Datacom from consulting firm Accenture Interactive and 44 00:02:50,800 --> 00:02:54,720 Speaker 3: has a background in user experience, design and publishing, has 45 00:02:54,760 --> 00:02:59,760 Speaker 3: seen firsthand how AI is amplifying rather than replacing human 46 00:02:59,760 --> 00:03:03,239 Speaker 3: crea activity. She'll also share the inside story of Datacom's 47 00:03:03,240 --> 00:03:06,800 Speaker 3: collaboration with the Melbourne Comedy Festival, where they built the 48 00:03:06,880 --> 00:03:11,720 Speaker 3: Funny Finder chatbot to help festival goers discover new acts. 49 00:03:11,880 --> 00:03:14,880 Speaker 3: If you're curious, cautious, or just trying to figure out 50 00:03:14,880 --> 00:03:18,040 Speaker 3: where AI fits in your creative world, you won't want 51 00:03:18,080 --> 00:03:31,720 Speaker 3: to miss this conversation. Here's my chat with Lou Compagnoni. Lou, 52 00:03:31,919 --> 00:03:33,799 Speaker 3: Welcome to the Business of Tech. How are you doing? 53 00:03:33,840 --> 00:03:35,280 Speaker 1: Thanks for having me. I'm doing well. 54 00:03:35,320 --> 00:03:36,280 Speaker 2: Thanks, how are you great? 55 00:03:36,280 --> 00:03:39,280 Speaker 3: It was great to see you in Wellington last week. 56 00:03:39,320 --> 00:03:44,600 Speaker 3: A stunning day at Datacom's Wellington headquarters panoramic views of 57 00:03:45,160 --> 00:03:49,200 Speaker 3: the Wellington Harbor. And that was the second of a 58 00:03:49,280 --> 00:03:53,640 Speaker 3: two part series that you hosted in New Zealand about 59 00:03:54,600 --> 00:03:58,720 Speaker 3: AI and the creative sector and how AI is being 60 00:03:58,760 --> 00:04:02,600 Speaker 3: adopted there, the opportunity and also the challenges that the 61 00:04:02,720 --> 00:04:05,400 Speaker 3: AI poses. So we're going to get into that, but 62 00:04:05,480 --> 00:04:09,000 Speaker 3: first a little bit about you. You have a creative 63 00:04:09,040 --> 00:04:12,040 Speaker 3: background as well. You did the same degree that I 64 00:04:12,080 --> 00:04:15,360 Speaker 3: did here in Wellington, the Institute of Modern Leaders Creative 65 00:04:15,360 --> 00:04:18,520 Speaker 3: Writing Masters. You were writing a novel. I was doing 66 00:04:18,640 --> 00:04:21,279 Speaker 3: a screenplay. Tell us a little bit about your experience 67 00:04:21,320 --> 00:04:21,640 Speaker 3: doing that. 68 00:04:21,920 --> 00:04:22,800 Speaker 1: Yeah, sure, So. 69 00:04:22,880 --> 00:04:25,600 Speaker 4: I actually decided to Having done a number of years 70 00:04:25,640 --> 00:04:28,279 Speaker 4: in the tech industry, I decided to go back to 71 00:04:28,320 --> 00:04:30,680 Speaker 4: my roots. My undergrad was in creative writing, and I 72 00:04:30,760 --> 00:04:32,360 Speaker 4: used to work as a food riser. 73 00:04:32,520 --> 00:04:35,880 Speaker 1: And I just wanted a chance to spend some time, just. 74 00:04:35,839 --> 00:04:38,400 Speaker 4: A dedicated time for eight months, really focusing on writing, 75 00:04:38,440 --> 00:04:40,520 Speaker 4: because I was finding it hard to dip in and 76 00:04:40,600 --> 00:04:44,640 Speaker 4: out of a day job in tech and writing. So yeah, 77 00:04:44,640 --> 00:04:46,279 Speaker 4: it was just really a chance to work on a 78 00:04:46,320 --> 00:04:48,760 Speaker 4: story that had been circling around in my brain and 79 00:04:48,800 --> 00:04:49,800 Speaker 4: needed some airtime. 80 00:04:50,320 --> 00:04:52,880 Speaker 1: It felt like a very luxurious. 81 00:04:52,279 --> 00:04:55,279 Speaker 4: Experience to spend so much time in the creative world, 82 00:04:55,680 --> 00:04:57,000 Speaker 4: and I'm really glad that I did it. 83 00:04:57,040 --> 00:04:59,000 Speaker 1: But like many people who did that course. 84 00:04:58,800 --> 00:05:02,400 Speaker 4: I still have this ninety percent done novel that I 85 00:05:02,480 --> 00:05:04,440 Speaker 4: keep wanting to return to. I don't know if you're 86 00:05:04,480 --> 00:05:06,679 Speaker 4: in the same situation with your script totally. 87 00:05:06,720 --> 00:05:10,839 Speaker 3: I still haven't produced my screenplay, so it's it's a 88 00:05:10,880 --> 00:05:14,400 Speaker 3: lifelong quest for some of us. But what an experience. 89 00:05:14,480 --> 00:05:18,120 Speaker 3: And I think really you could see that coming through 90 00:05:18,320 --> 00:05:22,919 Speaker 3: in Wellington, you know, the way that you've approached creativity 91 00:05:22,960 --> 00:05:26,040 Speaker 3: and you're a music maker as well, writing songs and 92 00:05:26,880 --> 00:05:33,480 Speaker 3: producing them to how this converging world is going to 93 00:05:33,839 --> 00:05:37,880 Speaker 3: have huge implications for people in the creative sector. You've 94 00:05:37,880 --> 00:05:41,240 Speaker 3: got a lot of people who are in the room 95 00:05:41,440 --> 00:05:46,560 Speaker 3: in Wellington from journalism outlets the likes of iron Z BBC. 96 00:05:47,880 --> 00:05:51,719 Speaker 3: You've got people there from game development companies like Pickpock, 97 00:05:52,640 --> 00:05:57,680 Speaker 3: and educational institutions at Massive University, the Creative Center at 98 00:05:57,520 --> 00:06:00,919 Speaker 3: a Mirramar which is right beside WET, a lot of 99 00:06:00,920 --> 00:06:04,720 Speaker 3: film students there. And the sense I got what they're 100 00:06:04,760 --> 00:06:09,080 Speaker 3: really saying all these great panelists was they see AI 101 00:06:09,160 --> 00:06:13,480 Speaker 3: as a tool to amplify human creativity and not replace it. 102 00:06:14,120 --> 00:06:18,200 Speaker 3: So you know, the students in Miramar are using AI, 103 00:06:19,000 --> 00:06:23,400 Speaker 3: but not to completely automate the process of creating a 104 00:06:23,440 --> 00:06:26,839 Speaker 3: short film or an animation. They're using it to maybe 105 00:06:26,960 --> 00:06:30,880 Speaker 3: iterate something to start out on that process. But they 106 00:06:30,920 --> 00:06:33,080 Speaker 3: actually want to be hands on, and it's not just 107 00:06:33,520 --> 00:06:35,960 Speaker 3: that they're being told to by their tutors saying you're 108 00:06:35,960 --> 00:06:38,440 Speaker 3: not allowed to use AI. They actually want to do that. 109 00:06:38,600 --> 00:06:39,960 Speaker 3: Is that the sense that you got as well? 110 00:06:40,040 --> 00:06:43,480 Speaker 4: Yeah, I definitely felt like I was surprised actually by 111 00:06:43,600 --> 00:06:46,960 Speaker 4: just how positive people's attitude was towards AI in that session. 112 00:06:47,000 --> 00:06:48,320 Speaker 4: I thought we were going to get a little bit 113 00:06:48,320 --> 00:06:52,320 Speaker 4: of a balance between excitement about it and experimentation, but 114 00:06:52,440 --> 00:06:54,800 Speaker 4: also a lot of fear and talks about copyright, but 115 00:06:54,880 --> 00:06:57,040 Speaker 4: I actually found that it was more on the innovation 116 00:06:57,200 --> 00:07:00,400 Speaker 4: side of things. I also had some interesting conversations in 117 00:07:00,440 --> 00:07:02,880 Speaker 4: the break, not just about how people are using it 118 00:07:02,920 --> 00:07:05,919 Speaker 4: for part of their creative process, but how they're using 119 00:07:05,920 --> 00:07:08,039 Speaker 4: it to try and free them up from the admin 120 00:07:08,120 --> 00:07:10,960 Speaker 4: tasks that take them away from their creative process. So, 121 00:07:11,040 --> 00:07:13,680 Speaker 4: for example, one of the photographers I knew there who 122 00:07:13,760 --> 00:07:17,600 Speaker 4: was actually photographing the event, Stephen, he was talking about 123 00:07:17,640 --> 00:07:20,680 Speaker 4: how he's actually working on developing a series of AI 124 00:07:20,760 --> 00:07:23,760 Speaker 4: agents to help him with his photography business, so it 125 00:07:23,800 --> 00:07:28,080 Speaker 4: can help with things like triaging emails, invoicing, sending out 126 00:07:29,400 --> 00:07:32,560 Speaker 4: JPEGs of photoshoots and so on, because he wants to 127 00:07:32,560 --> 00:07:35,000 Speaker 4: spend time making and so that makes so much sense. 128 00:07:36,640 --> 00:07:41,520 Speaker 3: Yeah, and Tyrone McCauley, one of the co founders of Pickpock, 129 00:07:41,600 --> 00:07:46,440 Speaker 3: the Great Wellington game development studio, he said they're really 130 00:07:46,560 --> 00:07:49,840 Speaker 3: using it to improve productivity, prototyping a game. You know, 131 00:07:49,920 --> 00:07:53,080 Speaker 3: that whole process has gone from weeks or days down 132 00:07:53,080 --> 00:07:59,840 Speaker 3: to hours. It's speeding up content creation, reducing repetitive coding work. 133 00:08:00,360 --> 00:08:05,040 Speaker 3: It's a way for them to explore new ideas. So 134 00:08:05,040 --> 00:08:08,800 Speaker 3: so that's I guess the story we've heard across industries 135 00:08:08,840 --> 00:08:13,360 Speaker 3: so far as that productivity story. That is the return 136 00:08:13,440 --> 00:08:17,120 Speaker 3: on investment for a lot of New Zealand and Australian businesses. 137 00:08:17,720 --> 00:08:21,040 Speaker 3: So that is starting to happen. The next phase is 138 00:08:21,080 --> 00:08:22,840 Speaker 3: going to be how do you use this to really 139 00:08:23,120 --> 00:08:28,000 Speaker 3: supercharge your business? And I was sitting next to a 140 00:08:28,040 --> 00:08:32,040 Speaker 3: folly artist, you know, who works with filmmakers to create 141 00:08:32,080 --> 00:08:35,960 Speaker 3: all the bang crash noises in the background. And that 142 00:08:36,160 --> 00:08:38,760 Speaker 3: is an art, you know, you're literally your timing has 143 00:08:38,800 --> 00:08:43,040 Speaker 3: to be impeccable, you have to have a really lifelike sound, 144 00:08:43,800 --> 00:08:47,320 Speaker 3: and there are libraries of those sorts of effects. But 145 00:08:47,559 --> 00:08:49,199 Speaker 3: I was sort of saying, something, what's going to happen 146 00:08:49,320 --> 00:08:52,800 Speaker 3: when when sort of AI can create some of these 147 00:08:52,800 --> 00:08:56,360 Speaker 3: sound effects instead of you in your folly studio banging 148 00:08:56,360 --> 00:08:58,400 Speaker 3: and crashing around. And he said, well, that's why I'm here. 149 00:08:58,800 --> 00:09:01,120 Speaker 3: I need to figure out what the future of my 150 00:09:01,440 --> 00:09:04,440 Speaker 3: particular niche in the film industry and he worked on 151 00:09:04,480 --> 00:09:07,559 Speaker 3: the Minecraft movie. You know, he's done some big productions 152 00:09:08,240 --> 00:09:11,920 Speaker 3: and since I got while there was positivity and optimism 153 00:09:11,920 --> 00:09:14,840 Speaker 3: in the room, there was also this sort of elephant 154 00:09:14,840 --> 00:09:16,560 Speaker 3: in the room as well as fear isn't going to 155 00:09:16,559 --> 00:09:20,760 Speaker 3: take my job. So I guess that is the big 156 00:09:20,840 --> 00:09:23,200 Speaker 3: question we You know, you talked about copyright, and I 157 00:09:23,200 --> 00:09:25,520 Speaker 3: think we sort of talked around that issue a little bit. 158 00:09:26,080 --> 00:09:29,120 Speaker 3: For instance, that Radio New Zealand was there, Glenn Scanlon, 159 00:09:29,200 --> 00:09:31,560 Speaker 3: who I was really interested to see what his take 160 00:09:31,640 --> 00:09:34,640 Speaker 3: on AI was, and he sort of said, well, we're 161 00:09:34,679 --> 00:09:37,000 Speaker 3: a public broadcaster. We like to give away all of 162 00:09:37,040 --> 00:09:40,440 Speaker 3: our content, so there's no point us actually trying to 163 00:09:40,480 --> 00:09:44,000 Speaker 3: block our content from being featured on chat, GPT or 164 00:09:44,120 --> 00:09:47,800 Speaker 3: perplexity because we want good quality information. 165 00:09:47,440 --> 00:09:48,200 Speaker 2: To get out there. 166 00:09:48,280 --> 00:09:51,520 Speaker 3: I guess the imperative is a little bit different for 167 00:09:51,600 --> 00:09:54,920 Speaker 3: those sorts of companies, you know, like my publisher in 168 00:09:55,080 --> 00:09:58,800 Speaker 3: znme that you know, the business model is very much 169 00:09:58,920 --> 00:10:03,440 Speaker 3: on copyright being maintained and being able to monetize that content. 170 00:10:03,679 --> 00:10:06,360 Speaker 4: Yeah, I think the big game changer in terms of 171 00:10:06,520 --> 00:10:09,280 Speaker 4: AI with that is really generative AI. So I think 172 00:10:09,280 --> 00:10:13,480 Speaker 4: there's a difference between consuming content and using that content 173 00:10:13,559 --> 00:10:16,360 Speaker 4: to create something new. And that's really a lot of 174 00:10:16,360 --> 00:10:19,200 Speaker 4: the concent that I've had from creatives that I know, 175 00:10:19,280 --> 00:10:23,000 Speaker 4: whether they're illustrators or writers, is there's a difference between 176 00:10:23,000 --> 00:10:25,120 Speaker 4: someone finding a link and looking at what they've made, 177 00:10:25,280 --> 00:10:28,920 Speaker 4: or reading their work and consuming their sort of style 178 00:10:29,120 --> 00:10:31,400 Speaker 4: to create something of their own. That's really where the 179 00:10:31,440 --> 00:10:32,440 Speaker 4: concerns are coming from. 180 00:10:32,760 --> 00:10:37,640 Speaker 3: Yeah, and there's no really good answer to that. Arguably 181 00:10:37,679 --> 00:10:39,320 Speaker 3: a lot of when it comes to journalism, a lot 182 00:10:39,360 --> 00:10:42,400 Speaker 3: of that's already been scraped, and there are some deals 183 00:10:42,679 --> 00:10:45,200 Speaker 3: being done. So if you can negotiate a deal with 184 00:10:45,280 --> 00:10:48,320 Speaker 3: the likes of open Ai, as some publishers have done, 185 00:10:48,679 --> 00:10:53,000 Speaker 3: that's a potential revenue stream. But in terms of creating 186 00:10:53,000 --> 00:10:56,080 Speaker 3: a movie or a song that is in the style 187 00:10:56,240 --> 00:11:00,280 Speaker 3: of a well known artist, that's a really tricky you on, 188 00:11:00,400 --> 00:11:04,040 Speaker 3: isn't it, Because where does that line blur between paying 189 00:11:04,120 --> 00:11:08,000 Speaker 3: homage to an artist that you like and literally copying 190 00:11:08,000 --> 00:11:10,600 Speaker 3: their style and in monetizing that yourself. 191 00:11:10,880 --> 00:11:12,960 Speaker 4: Often and it can be beyond your control, Like there's 192 00:11:12,960 --> 00:11:16,240 Speaker 4: even I saw a really interesting article recently with David Sidaris, 193 00:11:16,280 --> 00:11:18,480 Speaker 4: a writer who I really like, and he was saying 194 00:11:18,480 --> 00:11:22,440 Speaker 4: what happens if after I die, my niece decides to 195 00:11:22,640 --> 00:11:25,640 Speaker 4: upload all of my content into something and anyone can 196 00:11:25,679 --> 00:11:28,240 Speaker 4: write in my style. So I kind of wonder what 197 00:11:28,240 --> 00:11:30,760 Speaker 4: that's actually going to do to the creative process as well. 198 00:11:30,800 --> 00:11:33,959 Speaker 4: Will this actually stop people from creating freely or will 199 00:11:33,960 --> 00:11:36,880 Speaker 4: it stop people from sharing their content with the world. 200 00:11:36,960 --> 00:11:39,480 Speaker 4: I just think there's so many big questions around it. 201 00:11:39,640 --> 00:11:42,560 Speaker 3: One of the other interesting things that Tyrone from Pickpock 202 00:11:42,679 --> 00:11:46,920 Speaker 3: said when he talks to his staff about the you know, 203 00:11:46,960 --> 00:11:51,440 Speaker 3: the potential fears they have around AI is skill atrophy. 204 00:11:51,679 --> 00:11:55,599 Speaker 3: So there's this sense that particularly around coding, where you know, 205 00:11:56,000 --> 00:11:58,200 Speaker 3: we're seeing with the new version of Claude that just 206 00:11:58,200 --> 00:12:01,359 Speaker 3: came out this week and that apparently the coding capabilities 207 00:12:01,360 --> 00:12:03,960 Speaker 3: are just insane, you know. So we are seeing entry 208 00:12:04,040 --> 00:12:07,480 Speaker 3: level sort of coding jobs being automated at a very 209 00:12:07,600 --> 00:12:13,520 Speaker 3: rapid pace. So I guess there's concern that for people 210 00:12:13,679 --> 00:12:17,800 Speaker 3: entering the industry, what are they going to actually do 211 00:12:17,920 --> 00:12:20,360 Speaker 3: because some of those skills they're not going to need. 212 00:12:20,520 --> 00:12:23,720 Speaker 3: People further up who are a bit more experienced. A 213 00:12:23,720 --> 00:12:25,760 Speaker 3: lot of the stuff that is a big chunker off 214 00:12:25,800 --> 00:12:30,080 Speaker 3: their work suddenly might be no longer skills that are 215 00:12:30,120 --> 00:12:34,000 Speaker 3: in demand, so you're having to reskill and retrain constantly 216 00:12:34,040 --> 00:12:35,120 Speaker 3: now in the world of AI. 217 00:12:35,559 --> 00:12:38,240 Speaker 4: Yeah, I think there's also another issue in there, because 218 00:12:38,320 --> 00:12:40,720 Speaker 4: I agree with all of that, like will we actually 219 00:12:40,800 --> 00:12:42,920 Speaker 4: atrophy if we're not using the skills that we have 220 00:12:43,080 --> 00:12:44,280 Speaker 4: and we're kind of becoming. 221 00:12:44,000 --> 00:12:45,480 Speaker 1: More orchestrators of the tools. 222 00:12:45,679 --> 00:12:47,480 Speaker 4: Will we ever be able to go back to be 223 00:12:47,559 --> 00:12:49,760 Speaker 4: the actual producers of that thing we did in the 224 00:12:49,760 --> 00:12:52,400 Speaker 4: first place. But then I've also seen the flip side, 225 00:12:52,400 --> 00:12:54,959 Speaker 4: which I've seen with a lot of developers, where they 226 00:12:55,040 --> 00:12:57,400 Speaker 4: used to have an undulating day, so they used to 227 00:12:57,440 --> 00:12:59,599 Speaker 4: have sort of code that they could just do in 228 00:12:59,640 --> 00:13:01,320 Speaker 4: their slow and they used to just kick back and 229 00:13:01,360 --> 00:13:03,000 Speaker 4: put some tunes on, and then they would have sort 230 00:13:03,040 --> 00:13:05,800 Speaker 4: of more harder pieces of work, which is possibly thinking 231 00:13:05,800 --> 00:13:08,640 Speaker 4: about the overall architecture of something or sort of creating 232 00:13:08,679 --> 00:13:12,600 Speaker 4: something new. But now because they're using these tools, they 233 00:13:12,600 --> 00:13:15,520 Speaker 4: are always in this high value fit thinking space, And 234 00:13:15,559 --> 00:13:17,439 Speaker 4: what I've seen is actually a lot of burnout from 235 00:13:17,480 --> 00:13:18,920 Speaker 4: people because it's. 236 00:13:18,840 --> 00:13:19,680 Speaker 1: Almost the opposite. 237 00:13:19,760 --> 00:13:22,880 Speaker 4: Rather than their brain's atrophying, they're actually in this extremely 238 00:13:23,000 --> 00:13:26,760 Speaker 4: kind of high pressure orchestration mode all day because they're 239 00:13:26,760 --> 00:13:28,760 Speaker 4: doing five days of coding in two days. Like I 240 00:13:28,800 --> 00:13:31,040 Speaker 4: heard a story about one of our developers who did 241 00:13:31,040 --> 00:13:34,320 Speaker 4: something like eighty thousand lines of code and a weekend 242 00:13:34,480 --> 00:13:36,200 Speaker 4: or something like that. You know, this is just you're 243 00:13:36,360 --> 00:13:38,880 Speaker 4: in high productive mode. So I think that's another thing 244 00:13:38,920 --> 00:13:40,000 Speaker 4: to sort of bring to the table. 245 00:13:40,280 --> 00:13:44,640 Speaker 3: Wow, a lot of people were there from Central Government. 246 00:13:45,200 --> 00:13:48,120 Speaker 3: The government Chief Digital Officer was there, sort of laid 247 00:13:48,120 --> 00:13:52,920 Speaker 3: out where government is taking a I very much seemed experimentation, 248 00:13:53,520 --> 00:13:58,960 Speaker 3: tightly sandboxed sort of projects. Gov GPT, for instance, came 249 00:13:59,000 --> 00:14:03,200 Speaker 3: out at Callahan in a that was a project last year. 250 00:14:03,920 --> 00:14:07,439 Speaker 3: Actually the day that we're at that summit that evening, 251 00:14:07,480 --> 00:14:10,320 Speaker 3: they announced that they're going to work with AWS on 252 00:14:10,400 --> 00:14:13,199 Speaker 3: an all of government app for New Zealand, which sounds 253 00:14:13,200 --> 00:14:16,400 Speaker 3: pretty exciting very early days, but you know, AI may 254 00:14:16,440 --> 00:14:19,320 Speaker 3: get a look in as part of that. What was 255 00:14:19,320 --> 00:14:22,320 Speaker 3: your sense around where we're at in New Zealand when 256 00:14:22,360 --> 00:14:25,440 Speaker 3: it comes to public sector use of AI, and in 257 00:14:25,480 --> 00:14:29,560 Speaker 3: particular compared to Australia market, You know, so well. 258 00:14:29,560 --> 00:14:30,600 Speaker 1: I think it's mixed. 259 00:14:30,640 --> 00:14:33,400 Speaker 4: I think it really depends on the agency that you're 260 00:14:33,440 --> 00:14:36,760 Speaker 4: working with. I think that there's some that I've definitely seen. 261 00:14:36,560 --> 00:14:38,640 Speaker 1: And I'm not going to name any names for obvious reasons. 262 00:14:38,840 --> 00:14:41,080 Speaker 4: So some that I've seen that are actually doing a 263 00:14:41,080 --> 00:14:44,000 Speaker 4: lot of experimentation, They've got lots of proof of concepts. 264 00:14:44,360 --> 00:14:46,720 Speaker 4: They're now at that next phase where they're going, actually, 265 00:14:46,760 --> 00:14:49,400 Speaker 4: how do we take this further? How do we scale 266 00:14:49,440 --> 00:14:51,240 Speaker 4: what we're doing? How do we create a sort of 267 00:14:51,280 --> 00:14:55,040 Speaker 4: AI platform or repository so that we can have multiple 268 00:14:55,040 --> 00:14:57,560 Speaker 4: agents working across up business. So there's that side of things, 269 00:14:57,560 --> 00:14:59,960 Speaker 4: and they're looking at setting up a center of enablement 270 00:15:00,200 --> 00:15:02,840 Speaker 4: to really start to scale. But on the flip side, 271 00:15:02,840 --> 00:15:05,920 Speaker 4: there are some agencies who are really cautious and really 272 00:15:06,080 --> 00:15:09,120 Speaker 4: risk adverse, which is understandable, and they haven't even really 273 00:15:09,200 --> 00:15:11,480 Speaker 4: taken the first steps. They haven't even really switched on 274 00:15:12,040 --> 00:15:15,680 Speaker 4: co pilot within their M three six five environment, and 275 00:15:15,840 --> 00:15:16,960 Speaker 4: it's because they're sort of worried. 276 00:15:16,960 --> 00:15:17,600 Speaker 1: They don't want to. 277 00:15:17,520 --> 00:15:20,040 Speaker 4: Become front page news, they know that data is not 278 00:15:20,080 --> 00:15:21,720 Speaker 4: ready and so on. So I would say there's a 279 00:15:21,720 --> 00:15:24,880 Speaker 4: mixture in New Zealand for sure, and in Australia I 280 00:15:24,880 --> 00:15:25,400 Speaker 4: think the same. 281 00:15:25,480 --> 00:15:26,960 Speaker 1: Like certainly there are some. 282 00:15:26,880 --> 00:15:29,880 Speaker 4: Places in Australia that are sort of some organizations that 283 00:15:29,920 --> 00:15:32,720 Speaker 4: are well ahead, they've really been using AI for a 284 00:15:32,760 --> 00:15:36,120 Speaker 4: number of years. But there's also some who kind of 285 00:15:36,120 --> 00:15:38,640 Speaker 4: think that they're really ahead with it, but when you 286 00:15:38,720 --> 00:15:41,160 Speaker 4: kind of dig into it, they're really just experimenting with 287 00:15:41,200 --> 00:15:44,360 Speaker 4: one tool. And where I think that I'm not seeing 288 00:15:44,400 --> 00:15:48,320 Speaker 4: it is actually thinking more holistically about organizations, So not 289 00:15:48,400 --> 00:15:50,480 Speaker 4: just one tool here and there, but what could it 290 00:15:50,520 --> 00:15:52,520 Speaker 4: mean for them as a whole government organization. 291 00:15:52,680 --> 00:15:53,760 Speaker 1: What could they shift. 292 00:15:53,520 --> 00:15:55,400 Speaker 4: From and too as a result of AI, and what 293 00:15:55,440 --> 00:15:56,560 Speaker 4: does that mean for their future? 294 00:15:56,920 --> 00:15:59,480 Speaker 3: Yeah, someone from the government there come into debt. We're 295 00:15:59,480 --> 00:16:02,800 Speaker 3: really only to use you know, Microsoft Copilot at the moment, 296 00:16:02,800 --> 00:16:04,920 Speaker 3: that's the only approved tool that was corrected by the 297 00:16:06,000 --> 00:16:10,720 Speaker 3: government Chief Digital Officer. Any government agency has autonomy to 298 00:16:10,800 --> 00:16:13,560 Speaker 3: choose what AI tools that they want to use. But 299 00:16:13,600 --> 00:16:17,880 Speaker 3: I think it definitely showed that concern particularly about government 300 00:16:18,240 --> 00:16:21,720 Speaker 3: data going being fed into larger language models and potentially 301 00:16:22,440 --> 00:16:26,680 Speaker 3: exposing sensitive information. I know having talked to people that 302 00:16:26,760 --> 00:16:29,960 Speaker 3: that is a huge concern. But you know, they're laying 303 00:16:30,000 --> 00:16:33,400 Speaker 3: the groundworks. They've done the public sector guidance on use 304 00:16:33,440 --> 00:16:37,480 Speaker 3: of AI. They've got an AI roadmap for business and development. 305 00:16:37,760 --> 00:16:40,040 Speaker 3: Ministry of Culture and Heritage was there they were about 306 00:16:40,120 --> 00:16:43,360 Speaker 3: to do a consultation on use of AI in the 307 00:16:43,400 --> 00:16:45,840 Speaker 3: creative industry. So it feels like in New Zealand they're 308 00:16:45,840 --> 00:16:49,760 Speaker 3: sort of laying the groundwork, going through all of those 309 00:16:49,800 --> 00:16:54,680 Speaker 3: processes to make this a responsible approach to doing AI. 310 00:16:54,720 --> 00:16:57,040 Speaker 3: I guess the question is is when are we because 311 00:16:57,320 --> 00:16:59,720 Speaker 3: this is moving so quickly and other countries are adopting 312 00:16:59,720 --> 00:17:01,880 Speaker 3: it to really good effect, when are we going to 313 00:17:01,920 --> 00:17:05,120 Speaker 3: start accelerating our use of it? And I guess that's 314 00:17:05,160 --> 00:17:07,720 Speaker 3: where you come in with Datacom. You've got the confidence 315 00:17:07,760 --> 00:17:11,840 Speaker 3: and the people and the experience with twenty plus years 316 00:17:10,880 --> 00:17:13,800 Speaker 3: of developing AI to help them do it. 317 00:17:14,040 --> 00:17:14,639 Speaker 1: Yeah, certainly. 318 00:17:14,680 --> 00:17:16,760 Speaker 4: And I think that that's what we're trying to help 319 00:17:16,800 --> 00:17:21,080 Speaker 4: to balance within government organizations and all organizations. Really is 320 00:17:21,119 --> 00:17:25,280 Speaker 4: that balance between governance and having the right guard rails 321 00:17:25,320 --> 00:17:29,639 Speaker 4: in place, but also experimentation and innovation. And I actually 322 00:17:29,680 --> 00:17:31,720 Speaker 4: think it's possible to do both at once, and I 323 00:17:31,720 --> 00:17:34,199 Speaker 4: think that one can help the other. So, for example, 324 00:17:34,200 --> 00:17:37,240 Speaker 4: when we're helping people to think about governance develop policies, 325 00:17:38,720 --> 00:17:41,480 Speaker 4: I think actually is really helpful to also be developing 326 00:17:41,520 --> 00:17:43,960 Speaker 4: something creating a proof of concept at the same time, 327 00:17:44,160 --> 00:17:47,000 Speaker 4: because it's in creating that that you actually uncover some 328 00:17:47,080 --> 00:17:50,239 Speaker 4: of the challenges and blockers within your organization and you 329 00:17:50,240 --> 00:17:52,439 Speaker 4: start to you start to understand where the data is, 330 00:17:52,480 --> 00:17:55,680 Speaker 4: where some of the restrictions are, what other systems you're using, 331 00:17:55,720 --> 00:17:57,760 Speaker 4: and so on. So I think you're almost thinking about 332 00:17:57,800 --> 00:18:02,000 Speaker 4: having too parallel works. Streams within organizations can be. 333 00:18:01,960 --> 00:18:06,800 Speaker 3: Helpful, and there's a lot of organizations work with data 334 00:18:06,880 --> 00:18:10,160 Speaker 3: com But interestingly, you know, in the creative space, you've 335 00:18:10,160 --> 00:18:12,800 Speaker 3: worked on a project recently in Australia with a comedy 336 00:18:12,880 --> 00:18:18,199 Speaker 3: festival using AI to help attendees or people interested in 337 00:18:18,240 --> 00:18:21,040 Speaker 3: going to the comedy festival engage with the content figure 338 00:18:21,040 --> 00:18:25,280 Speaker 3: out exactly what shows they wanted to go to. 339 00:18:25,280 --> 00:18:26,439 Speaker 2: Took us through your. 340 00:18:26,320 --> 00:18:28,760 Speaker 3: Approach to that because that, potentially comedy could be quite 341 00:18:30,040 --> 00:18:32,879 Speaker 3: a touchy subject there. You know, we've had lots of 342 00:18:33,400 --> 00:18:37,480 Speaker 3: discussions about freedom of speech and what's appropriate and people 343 00:18:37,480 --> 00:18:40,160 Speaker 3: being canceled in comedy and all that sort of thing, 344 00:18:40,240 --> 00:18:43,000 Speaker 3: So putting AI into the mix sort of high risk. 345 00:18:43,440 --> 00:18:45,800 Speaker 4: It was, yeah, and we definitely had some moments, so 346 00:18:46,040 --> 00:18:48,359 Speaker 4: there was a lot of considerations with that. I guess 347 00:18:48,400 --> 00:18:51,880 Speaker 4: the reason it came about was the comedy festival. We're 348 00:18:51,880 --> 00:18:53,560 Speaker 4: getting to a point where I just had so many 349 00:18:53,600 --> 00:18:57,440 Speaker 4: events that even the big acts everyone was able to 350 00:18:57,480 --> 00:18:59,560 Speaker 4: find their comedy, but the sort of new and emerging 351 00:18:59,600 --> 00:19:02,760 Speaker 4: ones weren't necessarily getting the airtime that they required, and 352 00:19:02,760 --> 00:19:04,160 Speaker 4: people were just kind of going to all of. 353 00:19:04,080 --> 00:19:05,000 Speaker 1: The usual shows. 354 00:19:05,400 --> 00:19:07,560 Speaker 4: So they just wanted it as a way a different 355 00:19:07,600 --> 00:19:11,320 Speaker 4: way for people to explore what they could possibly see. 356 00:19:12,200 --> 00:19:14,440 Speaker 4: But we definitely had a lot of considerations along the way. 357 00:19:14,480 --> 00:19:17,479 Speaker 4: So one was to your point, you know, the nature 358 00:19:17,480 --> 00:19:20,280 Speaker 4: of comedy, it's not always the most PG content, So 359 00:19:20,320 --> 00:19:22,280 Speaker 4: what was it going to surface and how was that 360 00:19:22,320 --> 00:19:25,400 Speaker 4: going to reflect on both the Melbourne Comedy as a brand, 361 00:19:25,440 --> 00:19:28,399 Speaker 4: but also us Startacom as a brand as well. We 362 00:19:28,520 --> 00:19:31,879 Speaker 4: also really needed to consider the comedians, so we needed 363 00:19:31,880 --> 00:19:34,800 Speaker 4: to consider like were we giving sort of fair exposure 364 00:19:34,920 --> 00:19:37,480 Speaker 4: to all comedians or was it going to favor some 365 00:19:38,400 --> 00:19:41,000 Speaker 4: And we also had to be very explicit about the 366 00:19:41,080 --> 00:19:44,000 Speaker 4: role of the Funny Finder, what was it meant to 367 00:19:44,000 --> 00:19:46,920 Speaker 4: do and what was its persona And so we actually 368 00:19:46,920 --> 00:19:49,960 Speaker 4: did a really good co design process with comedians and 369 00:19:50,040 --> 00:19:53,000 Speaker 4: the Melbourne Comedy Festival and obviously the Data com team 370 00:19:53,359 --> 00:19:56,159 Speaker 4: to really explore that and we asked questions like, is 371 00:19:56,200 --> 00:19:58,200 Speaker 4: it funny, Should it be funny, should it be telling 372 00:19:58,280 --> 00:20:00,560 Speaker 4: jokes or should it actually just be really in formative? 373 00:20:00,640 --> 00:20:02,240 Speaker 1: Is it a comedian as a persona? 374 00:20:03,240 --> 00:20:05,679 Speaker 4: What we actually landed on was, I don't know if 375 00:20:05,680 --> 00:20:07,920 Speaker 4: you've been to the comedy show, but at the town 376 00:20:08,000 --> 00:20:10,080 Speaker 4: hall they have these sort of booths where you can 377 00:20:10,480 --> 00:20:12,399 Speaker 4: talk to people who know about the comedy show and 378 00:20:12,440 --> 00:20:15,560 Speaker 4: ask for advice about shows to see. They're really fun people, 379 00:20:15,560 --> 00:20:17,280 Speaker 4: They're really upbeat and really friendly. 380 00:20:17,720 --> 00:20:19,200 Speaker 1: So we decided that is. 381 00:20:19,160 --> 00:20:21,680 Speaker 4: Exactly the persona that we wanted it to be. They're 382 00:20:21,720 --> 00:20:25,760 Speaker 4: really nice, but they're not comedians, So we kept going 383 00:20:25,800 --> 00:20:26,480 Speaker 4: back to that thing. 384 00:20:26,600 --> 00:20:28,800 Speaker 1: The job of the Funny Finder is not to tell jokes. 385 00:20:29,280 --> 00:20:32,840 Speaker 4: However, given the nature of AI, there were moments when 386 00:20:32,840 --> 00:20:35,960 Speaker 4: the Funny Finder tried to become funny, and when we're 387 00:20:35,960 --> 00:20:38,679 Speaker 4: in the development and testing process sort of started to 388 00:20:39,400 --> 00:20:44,240 Speaker 4: become its own sort of comedian. So there was a 389 00:20:44,280 --> 00:20:49,000 Speaker 4: moment when we were doing some testing and it actually 390 00:20:49,040 --> 00:20:52,119 Speaker 4: started to make up comedy shows. This was obviously before release, 391 00:20:52,960 --> 00:20:54,879 Speaker 4: so we had to be very much going back to 392 00:20:54,920 --> 00:20:58,000 Speaker 4: that whole guideline of No, your job is to just 393 00:20:58,040 --> 00:21:00,600 Speaker 4: give information about the comedy show. We had to turn 394 00:21:00,680 --> 00:21:03,679 Speaker 4: the temperature down of the model and get it back 395 00:21:03,720 --> 00:21:07,159 Speaker 4: to its original intent, which is just to be informative. 396 00:21:07,200 --> 00:21:09,280 Speaker 4: So yeah, there was some really interesting issues to explore 397 00:21:09,320 --> 00:21:10,160 Speaker 4: on that project. 398 00:21:10,800 --> 00:21:13,840 Speaker 3: Wow, that's fascinating where it's literally starting to it's so 399 00:21:14,840 --> 00:21:18,000 Speaker 3: powerful and innovative in its own way. It's coming up 400 00:21:18,040 --> 00:21:20,840 Speaker 3: with comedy routines and I guess you're going to get 401 00:21:21,040 --> 00:21:24,240 Speaker 3: naturally enough people prompting it to do all sorts of 402 00:21:24,280 --> 00:21:26,680 Speaker 3: crazy things. So you know that is a real risk 403 00:21:26,760 --> 00:21:31,520 Speaker 3: of alienating and you know an audience and comedians go 404 00:21:31,600 --> 00:21:36,679 Speaker 3: into blurry territory around racial matter and sexism and all 405 00:21:36,680 --> 00:21:39,320 Speaker 3: that sort of thing, so potentially go wrong. But it 406 00:21:39,320 --> 00:21:41,120 Speaker 3: seems like it went really well. What was the sort 407 00:21:41,119 --> 00:21:44,520 Speaker 3: of engagement with the Funny Finder, Like, well. 408 00:21:44,400 --> 00:21:48,240 Speaker 4: We had around seven thousand queries per day. I mean, 409 00:21:48,280 --> 00:21:50,720 Speaker 4: given that this was actually quite early on, we thought 410 00:21:51,160 --> 00:21:53,359 Speaker 4: I think it was overall it was around four hundred 411 00:21:53,400 --> 00:21:56,320 Speaker 4: thousand queries that went through the Funny Finder, which we 412 00:21:56,320 --> 00:21:59,879 Speaker 4: thought was actually for a first year, quite good. We 413 00:22:00,080 --> 00:22:02,520 Speaker 4: also had really good visibility of how people were using it, 414 00:22:02,560 --> 00:22:04,000 Speaker 4: so we've got really good stats on what were the 415 00:22:04,080 --> 00:22:06,200 Speaker 4: kind of prompts that people were putting into it and 416 00:22:06,640 --> 00:22:09,359 Speaker 4: what was it offering in terms of information, So that 417 00:22:09,440 --> 00:22:12,200 Speaker 4: in itself is really interesting. Like we kind of expected 418 00:22:12,240 --> 00:22:13,840 Speaker 4: it would be more like I like this kind of 419 00:22:13,840 --> 00:22:16,560 Speaker 4: comedy and my friend likes this, so can you recommend 420 00:22:16,560 --> 00:22:18,760 Speaker 4: a show? But people were using it for all sorts 421 00:22:18,800 --> 00:22:23,480 Speaker 4: of things, including planning second dates, which was a surprise use. 422 00:22:24,200 --> 00:22:25,000 Speaker 2: That's fantastic. 423 00:22:25,040 --> 00:22:28,160 Speaker 3: So I guess you know that is that's a creative 424 00:22:28,200 --> 00:22:31,720 Speaker 3: sector example of that. But that sort of thing is 425 00:22:31,720 --> 00:22:34,960 Speaker 3: going across all the industries that you're working in, and 426 00:22:36,240 --> 00:22:37,880 Speaker 3: it's been going on for a while. It's getting better 427 00:22:37,920 --> 00:22:41,760 Speaker 3: and better the results, the information that you get from 428 00:22:42,200 --> 00:22:44,560 Speaker 3: these sorts of chatbots. I guess you know, as we 429 00:22:44,760 --> 00:22:48,919 Speaker 3: enter the agentic world, we're hearing so much about you know, 430 00:22:48,960 --> 00:22:51,119 Speaker 3: what's going to come next is you know, I'll be 431 00:22:51,160 --> 00:22:53,120 Speaker 3: able to tell that chatbot okay, I like the sound 432 00:22:53,160 --> 00:22:54,760 Speaker 3: of that. Book, me a couple of tickets and we'll 433 00:22:54,800 --> 00:22:56,520 Speaker 3: go off and do that on your behalf. 434 00:22:56,600 --> 00:22:57,320 Speaker 1: Yeah. Absolutely. 435 00:22:57,440 --> 00:22:59,439 Speaker 4: We even talked about that kind of thing as a 436 00:22:59,480 --> 00:23:01,840 Speaker 4: future faith So we kind of just wanted to deliver 437 00:23:02,000 --> 00:23:04,480 Speaker 4: the first year on just actually making sure it was safe, 438 00:23:04,480 --> 00:23:06,800 Speaker 4: that it didn't do anything Kooki that we weren't expecting, 439 00:23:07,480 --> 00:23:10,480 Speaker 4: and it just delivered on explaining the comedy show. But yeah, 440 00:23:10,480 --> 00:23:12,760 Speaker 4: in the future, we've talked about actually the ability for 441 00:23:12,800 --> 00:23:15,440 Speaker 4: it to become almost your travel planner, so it could 442 00:23:15,480 --> 00:23:17,720 Speaker 4: even have agents that go beyond booking your ticket. It 443 00:23:17,720 --> 00:23:20,439 Speaker 4: could also book you a restaurant that's nearby the comedy 444 00:23:20,480 --> 00:23:23,840 Speaker 4: show and audio an uber and just become sort of 445 00:23:23,840 --> 00:23:26,200 Speaker 4: your whole kind of concierge for the night. I definitely 446 00:23:26,240 --> 00:23:29,280 Speaker 4: think there's lots of ability to do that. It's just 447 00:23:29,320 --> 00:23:32,399 Speaker 4: a matter of I guess, investment and time and like 448 00:23:32,440 --> 00:23:34,439 Speaker 4: I said, just actually testing just to make sure that 449 00:23:34,480 --> 00:23:35,639 Speaker 4: everything is safe before you do. 450 00:23:35,720 --> 00:23:37,520 Speaker 3: So that's sort of sense when you look at the 451 00:23:37,560 --> 00:23:41,640 Speaker 3: creative sector as a whole, that because creativity is seen 452 00:23:41,680 --> 00:23:45,840 Speaker 3: as the secret source of that industry that you know, 453 00:23:46,080 --> 00:23:49,240 Speaker 3: they're sort of doing peripheral things with AI around productivity 454 00:23:49,720 --> 00:23:54,080 Speaker 3: and the like, speeding things up, but are they actually 455 00:23:54,280 --> 00:23:58,280 Speaker 3: using it for the core purpose of creating new and 456 00:23:58,400 --> 00:24:01,680 Speaker 3: novel things. Are you seeing examples across the tasma there 457 00:24:01,680 --> 00:24:07,359 Speaker 3: off creative agencies, maybe advertising agencies that are using AI, 458 00:24:07,600 --> 00:24:10,040 Speaker 3: if not for the actual creation of the end product 459 00:24:10,080 --> 00:24:13,200 Speaker 3: that's going out on radio or TV. But that really 460 00:24:13,240 --> 00:24:17,280 Speaker 3: difficult process of coming up with interesting ideas. Are they 461 00:24:17,359 --> 00:24:20,119 Speaker 3: augmenting that process with AI? 462 00:24:20,280 --> 00:24:22,440 Speaker 1: Yeah, I'm definitely seeing it sort of all throughout. 463 00:24:22,480 --> 00:24:26,120 Speaker 4: You know, as designers, we think of the double diamond process, 464 00:24:26,160 --> 00:24:28,399 Speaker 4: so there's the kind of like discovery and then the 465 00:24:28,480 --> 00:24:32,680 Speaker 4: define and then the design and deliver. So I'm definitely 466 00:24:32,680 --> 00:24:35,000 Speaker 4: seeing it all throughout. So I'm actually seeing people doing 467 00:24:35,000 --> 00:24:37,920 Speaker 4: it in the research process as well, So everything from 468 00:24:38,640 --> 00:24:42,360 Speaker 4: coming up with ideas to looking at doing a discovery process, 469 00:24:42,400 --> 00:24:44,679 Speaker 4: to working out where problems are that need to be 470 00:24:44,760 --> 00:24:48,359 Speaker 4: solved and there's opportunities, so using AI to help with 471 00:24:48,480 --> 00:24:51,920 Speaker 4: speeding up those sorts of processes as well as actually 472 00:24:52,040 --> 00:24:54,159 Speaker 4: using it as a tool to create as well. And 473 00:24:54,480 --> 00:24:56,040 Speaker 4: some of the people that I know who are creative 474 00:24:56,240 --> 00:24:58,480 Speaker 4: are using it for amazing things. So I know someone 475 00:24:58,480 --> 00:25:01,879 Speaker 4: who's a jewelry designer and she describes the way that 476 00:25:01,920 --> 00:25:05,560 Speaker 4: she designs jewelry now like creating with words or painting 477 00:25:05,640 --> 00:25:09,200 Speaker 4: with words to do some of her designs. So she's incredible. 478 00:25:09,240 --> 00:25:11,360 Speaker 4: If you weren't a jewelry designer, you wouldn't even know 479 00:25:11,440 --> 00:25:13,080 Speaker 4: how to come up with the prompts that she comes 480 00:25:13,119 --> 00:25:15,399 Speaker 4: up with. But she's shown me some of the designs 481 00:25:15,400 --> 00:25:17,160 Speaker 4: that she said just would merely would not have been 482 00:25:17,200 --> 00:25:20,400 Speaker 4: possible even to draw previously, that she's come up with. 483 00:25:21,200 --> 00:25:24,160 Speaker 4: I've also seen some interesting things with creatives where they're 484 00:25:24,200 --> 00:25:28,520 Speaker 4: actually giving creative voices to large language models, which is 485 00:25:28,520 --> 00:25:31,840 Speaker 4: a really cooky twist. So I mentioned doctor Michael Kollo 486 00:25:32,400 --> 00:25:34,840 Speaker 4: when I spoke at Creative Industries because he does a 487 00:25:34,880 --> 00:25:37,399 Speaker 4: lot of experiments with large language models, and one that 488 00:25:37,440 --> 00:25:39,480 Speaker 4: he sent me last night which was quite interesting, was 489 00:25:39,560 --> 00:25:42,480 Speaker 4: he got clawed to do a spoken word poem about 490 00:25:42,480 --> 00:25:43,520 Speaker 4: its own identity. 491 00:25:43,800 --> 00:25:45,080 Speaker 1: It was actually really beautiful. 492 00:25:45,119 --> 00:25:47,520 Speaker 4: It was quite disturbing at times, but I thought that 493 00:25:47,560 --> 00:25:50,080 Speaker 4: was a really interesting twist on creativity that you can 494 00:25:50,119 --> 00:25:54,400 Speaker 4: actually give voice to models themselves, so that opens all 495 00:25:54,440 --> 00:25:55,920 Speaker 4: sorts of new questions. 496 00:25:55,840 --> 00:26:00,879 Speaker 3: And that beautiful example that might have been You showed 497 00:26:01,400 --> 00:26:06,000 Speaker 3: a video off of the British Talco that invented the 498 00:26:06,040 --> 00:26:10,280 Speaker 3: AI granny to talk to scammers on the phone to 499 00:26:10,320 --> 00:26:11,119 Speaker 3: waste their time. 500 00:26:11,520 --> 00:26:12,160 Speaker 1: I love that. 501 00:26:12,400 --> 00:26:14,320 Speaker 4: I think that's brilliant because it is such a problem. 502 00:26:14,359 --> 00:26:16,600 Speaker 4: I know that that's something that my mum experiences a lot. 503 00:26:16,680 --> 00:26:18,679 Speaker 4: She gets lots of people scamming her, and it's really 504 00:26:18,760 --> 00:26:20,720 Speaker 4: difficult to discern for her whether it's. 505 00:26:20,560 --> 00:26:21,320 Speaker 1: A scam or not. 506 00:26:21,680 --> 00:26:23,720 Speaker 4: And I think it took a really creative person to 507 00:26:23,720 --> 00:26:25,760 Speaker 4: flip that on its head and to go, actually, we're 508 00:26:25,800 --> 00:26:29,480 Speaker 4: going to go after the scammers. Yeah, actually, And it 509 00:26:29,480 --> 00:26:31,720 Speaker 4: was quite funny. It was a playful way of dealing 510 00:26:31,760 --> 00:26:34,280 Speaker 4: with a problem. So I think that was one of 511 00:26:34,320 --> 00:26:37,359 Speaker 4: the things that I most liked talking about at AI 512 00:26:37,440 --> 00:26:40,359 Speaker 4: and creative industries, not just how you can use AI 513 00:26:41,080 --> 00:26:44,520 Speaker 4: as a creative tool, but also how creative mindsets are 514 00:26:44,600 --> 00:26:48,240 Speaker 4: needed to design AI and to think of new use 515 00:26:48,280 --> 00:26:51,000 Speaker 4: cases and innovations with AI, because I do think it 516 00:26:51,040 --> 00:26:52,560 Speaker 4: is such a different way of thinking. 517 00:26:52,760 --> 00:26:54,800 Speaker 1: Linear thinkers sometimes struggle with it. 518 00:26:55,160 --> 00:26:58,040 Speaker 3: A number of people commented there, and I think this, 519 00:26:58,400 --> 00:27:02,600 Speaker 3: in my experience, is so true. They see AI as 520 00:27:02,760 --> 00:27:08,160 Speaker 3: democratizing sort of some of the functions of creativity. And 521 00:27:08,560 --> 00:27:12,120 Speaker 3: you know, I've just finally just created a new website 522 00:27:12,119 --> 00:27:14,960 Speaker 3: for my own content thousands of articles over the years. 523 00:27:15,240 --> 00:27:17,320 Speaker 3: I've been putting it off for years because I hate 524 00:27:17,320 --> 00:27:24,000 Speaker 3: wrangling WordPress websites and getting into content management systems. And 525 00:27:24,680 --> 00:27:26,719 Speaker 3: I was just the other day, just waiting for a 526 00:27:26,760 --> 00:27:29,800 Speaker 3: conference to start. I went onto this app called Replet, 527 00:27:30,119 --> 00:27:35,919 Speaker 3: and it's an agent generator AI agent creator and was 528 00:27:35,960 --> 00:27:39,800 Speaker 3: able to create within two minutes just by prompting the agent. 529 00:27:39,880 --> 00:27:45,119 Speaker 3: There a website that does everything I need. It's just 530 00:27:45,160 --> 00:27:48,280 Speaker 3: a matter of porting over all of my content. And 531 00:27:48,440 --> 00:27:50,720 Speaker 3: apparently i'm reading up replet can do a bit of 532 00:27:50,720 --> 00:27:54,159 Speaker 3: code to actually automate that process as well. That just 533 00:27:54,200 --> 00:27:58,000 Speaker 3: blew my mind. So that democratization, if not off the 534 00:27:58,840 --> 00:28:02,480 Speaker 3: kernel of creativity can all be great novelists and songwriters, 535 00:28:02,880 --> 00:28:06,720 Speaker 3: but all the stuff that goes around it and allows 536 00:28:06,760 --> 00:28:09,000 Speaker 3: people to be more creative than they otherwise would be. 537 00:28:09,320 --> 00:28:10,280 Speaker 1: Yeah, I definitely think. 538 00:28:10,320 --> 00:28:12,080 Speaker 4: I mean, that's a really good side that I think 539 00:28:12,200 --> 00:28:15,760 Speaker 4: is coming out of this is really democratizing creativity. And 540 00:28:15,800 --> 00:28:18,200 Speaker 4: I think that's particularly good for small businesses who may 541 00:28:18,240 --> 00:28:21,200 Speaker 4: not have the budget to get professionals to do everything, 542 00:28:21,800 --> 00:28:26,040 Speaker 4: and they're able to get professional quality materials without hiring 543 00:28:26,119 --> 00:28:30,040 Speaker 4: a professional designer or a photographer or web developer. I 544 00:28:30,080 --> 00:28:32,679 Speaker 4: think what's interesting about that is what that means for 545 00:28:32,720 --> 00:28:35,720 Speaker 4: the role of creatives who are otherwise doing that work though, 546 00:28:36,000 --> 00:28:37,920 Speaker 4: And what I'm saying is a bit of a shift 547 00:28:38,080 --> 00:28:41,160 Speaker 4: in the way creative people are working and the new 548 00:28:41,240 --> 00:28:43,880 Speaker 4: roles that are emerging. So some people who are more 549 00:28:44,600 --> 00:28:48,440 Speaker 4: makers and creating things themselves are actually moving more into 550 00:28:48,440 --> 00:28:51,720 Speaker 4: creative direction. And similar to what I talked about with 551 00:28:51,760 --> 00:28:56,880 Speaker 4: developers becoming orchestrating the tools themselves, they're actually doing that. 552 00:28:56,920 --> 00:29:01,600 Speaker 4: So they're more doing content, actually using AI to generate 553 00:29:01,680 --> 00:29:04,080 Speaker 4: content themselves to do the editing, and they've been quite 554 00:29:04,120 --> 00:29:05,120 Speaker 4: transparent about the. 555 00:29:05,040 --> 00:29:06,840 Speaker 1: Fact that they're using it in their processes. 556 00:29:07,440 --> 00:29:09,240 Speaker 4: I'm also seeing people who are sort of moving into 557 00:29:09,320 --> 00:29:12,720 Speaker 4: roles like AI art directors, and I've heard this new 558 00:29:12,800 --> 00:29:14,800 Speaker 4: term recently. I'm not sure if you've heard this, called 559 00:29:14,800 --> 00:29:18,479 Speaker 4: a promptographer. That's a bit of a tongue twister, isn't it, 560 00:29:19,280 --> 00:29:21,760 Speaker 4: And so that emerged from there was a German artist 561 00:29:22,400 --> 00:29:27,160 Speaker 4: who entered a Sony photography exhibition and he actually entered 562 00:29:27,400 --> 00:29:31,600 Speaker 4: a AI generated photograph and the reason he did it 563 00:29:31,640 --> 00:29:34,120 Speaker 4: was not to try and sneakily win. So he won 564 00:29:34,120 --> 00:29:37,440 Speaker 4: the competition and when he was accepting the award, he 565 00:29:37,520 --> 00:29:40,600 Speaker 4: actually was transparent and said, actually, I will not take 566 00:29:40,640 --> 00:29:41,640 Speaker 4: this award. 567 00:29:41,720 --> 00:29:42,440 Speaker 1: I used AI. 568 00:29:43,040 --> 00:29:45,400 Speaker 4: There's a role for promptography, but we really need to 569 00:29:45,440 --> 00:29:48,720 Speaker 4: think about actually how we are transparent about what's human 570 00:29:48,720 --> 00:29:51,720 Speaker 4: made and what AI made in these processes. So I 571 00:29:51,760 --> 00:29:54,960 Speaker 4: think interesting lots of things around that democratization, but also 572 00:29:55,000 --> 00:29:57,160 Speaker 4: the impact of that on the creative industry. 573 00:29:57,520 --> 00:30:00,800 Speaker 3: And just finally, Leu, I think you were probably talking 574 00:30:00,800 --> 00:30:03,040 Speaker 3: to and I was there, and a number of people 575 00:30:03,320 --> 00:30:06,040 Speaker 3: who are at these conferences sort of sitting at the fence. 576 00:30:06,080 --> 00:30:09,400 Speaker 3: They know they need to upskill around AI and figure 577 00:30:09,440 --> 00:30:12,560 Speaker 3: out what it means for their business. They haven't really 578 00:30:12,600 --> 00:30:16,560 Speaker 3: progressed any further yet, which sort of surprised me. In 579 00:30:16,600 --> 00:30:19,080 Speaker 3: the creative sector, I thought there would be a lot 580 00:30:19,120 --> 00:30:23,080 Speaker 3: more sort of jumping in feet first, what do they 581 00:30:23,320 --> 00:30:25,880 Speaker 3: need to do and what can they do to sort 582 00:30:25,880 --> 00:30:29,680 Speaker 3: of get that at least get that experimentation going, while 583 00:30:29,720 --> 00:30:33,520 Speaker 3: they're also looking at the potential to transform their business, 584 00:30:33,560 --> 00:30:35,680 Speaker 3: which is what we really want. That's where the productivity 585 00:30:35,680 --> 00:30:38,320 Speaker 3: gains and the new features and products are going to 586 00:30:38,360 --> 00:30:43,440 Speaker 3: come from. How do they sort of approach that with 587 00:30:43,800 --> 00:30:45,280 Speaker 3: keeping in mind it's got to be a return on 588 00:30:45,360 --> 00:30:47,800 Speaker 3: investment on this. They can't just be buying widgets and 589 00:30:48,280 --> 00:30:51,080 Speaker 3: subscriptions and it's not going anywhere that it actually needs 590 00:30:51,120 --> 00:30:52,040 Speaker 3: to stack up financially. 591 00:30:52,160 --> 00:30:54,560 Speaker 4: I think there's almost two things to be thinking about 592 00:30:54,640 --> 00:30:57,360 Speaker 4: there's probably more than two to come to mind. So 593 00:30:57,480 --> 00:31:01,080 Speaker 4: one is just to start experimenting now rather than later. 594 00:31:01,280 --> 00:31:03,760 Speaker 4: So rather than going I'm waiting for the perfect AI 595 00:31:03,880 --> 00:31:06,880 Speaker 4: tool to help me do this, actually just to dive 596 00:31:06,920 --> 00:31:09,480 Speaker 4: into the current tools that are available and just start 597 00:31:09,480 --> 00:31:11,760 Speaker 4: playing around with what they can do. So you know, 598 00:31:11,840 --> 00:31:14,440 Speaker 4: things like if you're a designer, you might be thinking 599 00:31:14,440 --> 00:31:16,680 Speaker 4: about things like mid Journey, or then there's the kind 600 00:31:16,720 --> 00:31:19,719 Speaker 4: of big Swiss army knife ones like chat GPT, Gemini 601 00:31:19,800 --> 00:31:22,360 Speaker 4: is pretty amazing and Claude as well, and I think 602 00:31:22,440 --> 00:31:25,800 Speaker 4: just spend time playing with these tools every week, both 603 00:31:25,960 --> 00:31:29,280 Speaker 4: in the sort of admin things that they do as 604 00:31:29,400 --> 00:31:32,880 Speaker 4: creatives to try and make it commercial, but also even 605 00:31:32,920 --> 00:31:35,880 Speaker 4: in their own creativity just to see what's possible and 606 00:31:35,920 --> 00:31:38,760 Speaker 4: whether there is an opportunity for them to amplify what 607 00:31:38,800 --> 00:31:39,160 Speaker 4: they do. 608 00:31:39,880 --> 00:31:41,360 Speaker 1: But then I also think it's. 609 00:31:41,240 --> 00:31:43,960 Speaker 4: Important to look more at a future horizon as well, 610 00:31:44,360 --> 00:31:46,840 Speaker 4: and to sort of stay across what actually is changing 611 00:31:46,880 --> 00:31:49,080 Speaker 4: with AI and what they could shift from and to 612 00:31:49,240 --> 00:31:52,560 Speaker 4: as a result of AI. So I think that goes 613 00:31:52,560 --> 00:31:54,880 Speaker 4: to your point of actually how might it transform them, 614 00:31:55,200 --> 00:31:57,160 Speaker 4: and seat of starting from that point of well, this 615 00:31:57,240 --> 00:31:59,120 Speaker 4: is what an ideal future would look like for me, 616 00:31:59,160 --> 00:32:02,240 Speaker 4: this is how it might transform me, and then go 617 00:32:02,360 --> 00:32:02,680 Speaker 4: back to. 618 00:32:02,680 --> 00:32:03,920 Speaker 1: Actually like how do I get there? 619 00:32:04,000 --> 00:32:05,640 Speaker 4: If it is going to change, I am going to 620 00:32:05,920 --> 00:32:08,560 Speaker 4: completely pivot what I do now? What do I need 621 00:32:08,600 --> 00:32:11,600 Speaker 4: to do to get in to protect myself really to 622 00:32:12,000 --> 00:32:14,240 Speaker 4: set that in motion that ideal future. 623 00:32:14,360 --> 00:32:18,280 Speaker 3: Excellent good advice there, Lou and hey, thanks for hosting 624 00:32:18,880 --> 00:32:20,120 Speaker 3: those sessions. 625 00:32:20,160 --> 00:32:20,560 Speaker 1: That was. 626 00:32:22,160 --> 00:32:23,560 Speaker 2: Really thought provoking, I think. 627 00:32:23,600 --> 00:32:27,920 Speaker 3: So hopefully we'll see our creative sector adopts dot really 628 00:32:27,920 --> 00:32:31,240 Speaker 3: get that experimenting going, and I'll put a link to 629 00:32:31,560 --> 00:32:32,040 Speaker 3: It's funny. 630 00:32:32,040 --> 00:32:33,080 Speaker 2: Fine, it's still going. 631 00:32:33,240 --> 00:32:33,600 Speaker 1: It's not. 632 00:32:33,640 --> 00:32:35,520 Speaker 4: Actually we took it down at the end of the 633 00:32:35,680 --> 00:32:38,760 Speaker 4: Comedy Festival, but we may well do the same thing 634 00:32:38,880 --> 00:32:40,480 Speaker 4: next year. Not the same thing. We might even do 635 00:32:40,520 --> 00:32:42,360 Speaker 4: a slightly different thing next year. Who knows how. 636 00:32:42,320 --> 00:32:44,200 Speaker 1: Much the technology will have moved on then. 637 00:32:45,080 --> 00:32:47,400 Speaker 4: So yeah, not available too, But we are working on 638 00:32:47,480 --> 00:32:50,680 Speaker 4: a lot of other sort of conversational AI tools for customers. 639 00:32:50,720 --> 00:32:52,240 Speaker 4: That I mean is a use case that could be 640 00:32:52,320 --> 00:32:54,000 Speaker 4: used for so many different things as well. 641 00:32:54,080 --> 00:32:57,280 Speaker 3: So absolutely, and I haven't seen anything like that for 642 00:32:57,520 --> 00:33:00,280 Speaker 3: a festival on this side of the TASM bit they 643 00:33:00,320 --> 00:33:02,520 Speaker 3: festival or the film festival or whatever. So I think 644 00:33:03,000 --> 00:33:05,480 Speaker 3: that whole space how you engage with and try and 645 00:33:05,480 --> 00:33:07,880 Speaker 3: get an audience convert them to buying a ticket that's 646 00:33:07,920 --> 00:33:10,920 Speaker 3: about to be revolutionized with the use of AI. Thanks Lou, 647 00:33:11,160 --> 00:33:13,080 Speaker 3: thanks so much for coming on the Business of Tech. 648 00:33:13,120 --> 00:33:14,680 Speaker 3: Great to see you, and we'll talk again. 649 00:33:14,760 --> 00:33:16,520 Speaker 1: Great to see you too, Peter, thanks so much. 650 00:33:22,240 --> 00:33:23,840 Speaker 2: That's a wrap for today's episode. 651 00:33:23,840 --> 00:33:26,640 Speaker 3: If the Business of Tech are one hundredth episode, thanks 652 00:33:26,680 --> 00:33:30,920 Speaker 3: so much for tuning in and supporting the podcast over 653 00:33:31,320 --> 00:33:34,760 Speaker 3: three seasons. Off the Business of Tech really appreciate it. 654 00:33:34,960 --> 00:33:39,200 Speaker 3: Huge Thanks to Lou Compagnoni for sharing her insights and 655 00:33:39,240 --> 00:33:42,720 Speaker 3: stories from the creative optimism she's seeing on both sides 656 00:33:42,720 --> 00:33:45,960 Speaker 3: of the Tasman to the behind the scenes challenges of 657 00:33:46,080 --> 00:33:50,160 Speaker 3: building an AI for comedy fans. I think her advice 658 00:33:50,320 --> 00:33:53,800 Speaker 3: is pretty clear, don't wait for the perfect toolstart experimenting 659 00:33:53,960 --> 00:33:57,760 Speaker 3: now and imagine how AI could transform your work. I'm 660 00:33:57,760 --> 00:34:02,160 Speaker 3: definitely doing that. But at the Wellington AI conference, at 661 00:34:02,480 --> 00:34:05,480 Speaker 3: the cautious approach people are taking to using AI, which 662 00:34:05,520 --> 00:34:08,319 Speaker 3: I think says a couple of things. You know, there's 663 00:34:08,400 --> 00:34:14,160 Speaker 3: real genuine concern over quality with hallucinations, odd looking images, 664 00:34:14,520 --> 00:34:18,520 Speaker 3: you know, frankly AI slop where everything looks mass manufactured 665 00:34:18,520 --> 00:34:21,080 Speaker 3: and generic. You know, our creatives do not want to 666 00:34:21,120 --> 00:34:24,040 Speaker 3: be part of that. They're very sensitive to the feedback 667 00:34:24,040 --> 00:34:26,879 Speaker 3: from customers, which is that they don't really want AI 668 00:34:27,000 --> 00:34:31,000 Speaker 3: generated content. They value the human touch, so that's fair enough. 669 00:34:31,280 --> 00:34:35,200 Speaker 3: On the other hand, the using AI in their processes 670 00:34:35,480 --> 00:34:39,480 Speaker 3: to become more efficient and to boost productivity, which is 671 00:34:39,520 --> 00:34:41,960 Speaker 3: a good thing I think is the quality of AI 672 00:34:42,080 --> 00:34:46,359 Speaker 3: improves and as competitors here and abroad adopt it, the tech, 673 00:34:46,440 --> 00:34:50,520 Speaker 3: which is improving at an accelerating rate, will play a 674 00:34:50,560 --> 00:34:55,320 Speaker 3: greater role in the creation aspects of advertising, video making, 675 00:34:55,600 --> 00:34:59,480 Speaker 3: music making, and the like. That clearly has huge implications 676 00:34:59,480 --> 00:35:02,960 Speaker 3: for artists, for journalists, and anyone in the creative industries. 677 00:35:03,160 --> 00:35:05,680 Speaker 3: If you enjoyed this episode, don't forget to subscribe and 678 00:35:05,800 --> 00:35:09,319 Speaker 3: leave us a review on iHeartRadio or wherever you get 679 00:35:09,320 --> 00:35:12,960 Speaker 3: your podcasts. I'll be back next Thursday with another episode 680 00:35:13,000 --> 00:35:16,280 Speaker 3: of the Business of Tech. Until next time, keep exploring 681 00:35:16,560 --> 00:35:17,480 Speaker 3: and keep creating.