1 00:00:00,040 --> 00:00:03,640 Coming up today on Mini Tech Scroll, Meta introduces and 2 00:00:03,640 --> 00:00:07,360 backtracks on a new AI feature. How mainstream social media 3 00:00:07,360 --> 00:00:11,840 platforms are promoting nudify tools and character AI's foray 4 00:00:11,840 --> 00:00:14,800 into micro dramas. Hey everyone, and welcome to 5 00:00:14,800 --> 00:00:17,840 Mini Scroll, the daily Internet culture podcast bringing you the 6 00:00:17,840 --> 00:00:20,920 biggest stories from social media, the creator economy and 7 00:00:20,920 --> 00:00:24,000 the digital space every Monday through Thursday and sometimes 8 00:00:24,000 --> 00:00:26,080 Fridays. My name is Annabelle and I'm 9 00:00:26,080 --> 00:00:29,280 your Wednesday Mini Tech Scroll host and also the deputy editor 10 00:00:29,280 --> 00:00:32,920 of Australian in youth digital media brand Centennial World. 11 00:00:33,040 --> 00:00:36,280 Welcome back to Maytag Scroll for another Wednesday. 12 00:00:36,560 --> 00:00:39,920 I don't really think I have too many updates at the top of this 13 00:00:39,920 --> 00:00:43,000 week's episode, apart from the fact that probably the biggest 14 00:00:43,240 --> 00:00:47,720 story in tech this week was actually covered by Lauren on 15 00:00:47,720 --> 00:00:49,520 Monday. If you want to hear everything 16 00:00:49,520 --> 00:00:52,880 that has been going on with Thea, which is the platform that 17 00:00:52,880 --> 00:00:56,120 Phoebe Gates Co founded and how they are currently in the midst 18 00:00:56,160 --> 00:00:59,840 of a scandal and possible fraud, definitely check that episode 19 00:00:59,840 --> 00:01:01,440 out. We thought it would be better to 20 00:01:01,440 --> 00:01:04,680 cover it there since it was like super timely, but yes, other 21 00:01:04,680 --> 00:01:07,400 than that, Please remember to rate review and subscribe. 22 00:01:07,480 --> 00:01:09,880 OK, let's get into the content for this week. 23 00:01:09,880 --> 00:01:13,720 So we first really need to talk about everything that went on 24 00:01:13,720 --> 00:01:16,440 with Instagram and muse image last week. 25 00:01:16,440 --> 00:01:19,840 This story really hit the headlines on Thursday last week, 26 00:01:19,840 --> 00:01:22,080 and I was actually so annoyed that I didn't get a chance to 27 00:01:22,080 --> 00:01:24,400 cover it on mini tech. But there have also been some 28 00:01:24,400 --> 00:01:27,000 updates since then. So I am actually glad that I 29 00:01:27,000 --> 00:01:30,280 didn't cover it last week. So in case you have haven't 30 00:01:30,280 --> 00:01:33,880 heard anything about this, about a week ago Meta announced Muse 31 00:01:33,880 --> 00:01:37,280 Image, which is the first media generation model developed by 32 00:01:37,280 --> 00:01:40,600 Meta Superintelligence Labs, which is the AI division of 33 00:01:40,600 --> 00:01:43,040 Meta. So Muse Image, according to a 34 00:01:43,040 --> 00:01:46,280 blog post was described as quote our most advanced image 35 00:01:46,280 --> 00:01:49,120 generation model yet. But one of the capabilities that 36 00:01:49,120 --> 00:01:52,280 had been announced alongside Muse Image soon became the 37 00:01:52,280 --> 00:01:55,800 centre of controversy and it was absolutely everywhere. 38 00:01:55,840 --> 00:01:59,600 Basically, this feature allowed Meta users to create AI 39 00:01:59,600 --> 00:02:03,800 generated pictures of public profiles from Instagram and they 40 00:02:03,800 --> 00:02:07,720 could do this by simply tagging them in a Meta AI prompt. 41 00:02:07,800 --> 00:02:10,000 I should know that private accounts and accounts belonging 42 00:02:10,000 --> 00:02:12,720 to mine is what automatically excluded from the feature, but 43 00:02:12,720 --> 00:02:16,200 the rest of us had to opt out. They tried to sell this feature 44 00:02:16,200 --> 00:02:19,560 to users in the blog post by saying that Muse image was 45 00:02:19,640 --> 00:02:22,880 rooted in your own world and honestly, that positioning 46 00:02:22,880 --> 00:02:25,440 didn't really work. As the blog post reads quote, 47 00:02:25,440 --> 00:02:28,760 you can at mention Instagram accounts in the Meta AI app to 48 00:02:28,760 --> 00:02:31,920 bring specific Instagram profiles right into your images. 49 00:02:31,920 --> 00:02:34,840 Whether you want to design a custom event invitation, mock up 50 00:02:34,840 --> 00:02:37,160 a collaborative creative concept, or generate A 51 00:02:37,160 --> 00:02:41,040 personalized graphic tagging a username, let's Meta AI use 52 00:02:41,040 --> 00:02:44,000 public photos to build a visual that's ready to post. 53 00:02:44,160 --> 00:02:47,000 You have control over how your content can be tagged for AI 54 00:02:47,000 --> 00:02:50,160 creation with an easy setting to turn off this feature at any 55 00:02:50,160 --> 00:02:52,440 time. While there is an option to turn 56 00:02:52,440 --> 00:02:55,360 off the feature, as they noted in the blog post, there are so 57 00:02:55,360 --> 00:02:57,320 many issues with a feature like this. 58 00:02:57,320 --> 00:03:00,840 First and foremost is the fact that it is opt out as opposed to 59 00:03:00,960 --> 00:03:03,080 opt in. Considering there would be a 60 00:03:03,080 --> 00:03:06,200 minimal control over how your likeness could be used. 61 00:03:06,320 --> 00:03:09,600 It really feels like users should have to actively agree to 62 00:03:09,600 --> 00:03:13,280 this, and the fact that Meta did opt out kind of suggests to me 63 00:03:13,280 --> 00:03:15,600 that they were trying to force people to engage with this 64 00:03:15,600 --> 00:03:19,280 feature in one way or another. Another issue of course is that 65 00:03:19,280 --> 00:03:22,320 Instagram users could not give their consent to the AI 66 00:03:22,320 --> 00:03:25,360 generated content of them that would be using their likeness. 67 00:03:25,720 --> 00:03:28,480 Users who hadn't heard about this update would firstly have 68 00:03:28,480 --> 00:03:33,840 no idea that it exists and even still have their content used by 69 00:03:34,040 --> 00:03:36,920 AI, and they also wouldn't have been notified when someone 70 00:03:36,920 --> 00:03:39,080 reused it. This is obviously a major 71 00:03:39,080 --> 00:03:42,680 privacy threat and it really opens the door to potential 72 00:03:42,720 --> 00:03:46,240 harassment and impersonation for any public account. 73 00:03:46,240 --> 00:03:48,480 And it also just kind of feels like quite a jump. 74 00:03:48,480 --> 00:03:51,680 Like just because someone posts their pictures publicly, it 75 00:03:51,680 --> 00:03:54,280 doesn't mean that they should have their content used like 76 00:03:54,280 --> 00:03:55,600 that. Like obviously in this day and 77 00:03:55,840 --> 00:03:59,200 everyone is aware that that is a possibility, but a platform 78 00:03:59,200 --> 00:04:02,080 making it that accessible is another level. 79 00:04:02,080 --> 00:04:05,000 When I first started hearing about all this it definitely 80 00:04:05,000 --> 00:04:08,720 reminded me of Soras cameo feature from Open AI, but 81 00:04:08,720 --> 00:04:13,040 somehow this was even worse. If you guys don't remember, the 82 00:04:13,040 --> 00:04:16,160 cameo feature basically allowed users to quote UN quote drop 83 00:04:16,160 --> 00:04:20,040 themselves into any Sora generated scene, but users also 84 00:04:20,040 --> 00:04:23,640 were in control of their likeness to some degree as it 85 00:04:23,640 --> 00:04:26,040 was written in the Open AI blog post at the time. 86 00:04:26,040 --> 00:04:29,600 Quote with cameos you are in control of your likeness end to 87 00:04:29,600 --> 00:04:31,160 end with Sora. Really. 88 00:04:31,160 --> 00:04:34,480 You decide who can use your Cameo, and you can revoke access 89 00:04:34,480 --> 00:04:37,640 or remove any video that includes it at any time. 90 00:04:38,040 --> 00:04:41,200 Videos containing cameos of you, including drafts created by 91 00:04:41,200 --> 00:04:44,320 other people, are viewable by you at any time. 92 00:04:44,320 --> 00:04:46,960 Even though there seemed to be more tools in place when it came 93 00:04:46,960 --> 00:04:50,160 to Cameo, people were obviously still very stressed about this 94 00:04:50,160 --> 00:04:53,800 feature before Sora shut down. Obviously in saying that, there 95 00:04:53,800 --> 00:04:57,560 seemed to be even less oversight with the Muse image feature. 96 00:04:58,000 --> 00:05:01,080 Quickly, publications, creators and even parliamentarians here 97 00:05:01,080 --> 00:05:03,760 in Australia started posting information about this new 98 00:05:03,760 --> 00:05:07,120 feature and how to opt out of it if you are a public account. 99 00:05:07,120 --> 00:05:10,000 SAG AFTRA was one pretty vocal critic. 100 00:05:10,040 --> 00:05:12,720 The union issued a statement recommending that their members 101 00:05:12,720 --> 00:05:16,520 turn off this feature, but also questioning the opt out element 102 00:05:16,560 --> 00:05:19,640 as their statement read quote. Anything other than a clear and 103 00:05:19,640 --> 00:05:23,400 conspicuous opt in for these types of uses of Instagram users 104 00:05:23,400 --> 00:05:27,200 accounts is unacceptable and an utter miscalculation of public 105 00:05:27,200 --> 00:05:30,600 sentiment regarding the obvious dangers and harms inherent in 106 00:05:30,600 --> 00:05:32,920 such use. After days of this mounting 107 00:05:32,920 --> 00:05:35,640 backlash, Meta has since pulled the feature. 108 00:05:35,720 --> 00:05:38,760 They issued an edit to the blog post announcing news image, 109 00:05:38,760 --> 00:05:42,040 including an update on July 10. This is what I read quote. 110 00:05:42,080 --> 00:05:44,840 Earlier this week, we announced that one way for people to 111 00:05:44,840 --> 00:05:48,960 generate images in Meta AI is by at mentioning public Instagram 112 00:05:48,960 --> 00:05:50,760 accounts that they want to reference. 113 00:05:50,800 --> 00:05:54,040 Our intent was to provide a useful creative tool and to give 114 00:05:54,040 --> 00:05:56,960 people control over whether their public content could be 115 00:05:56,960 --> 00:05:59,640 referenced in this way. We've had the feedback that this 116 00:05:59,640 --> 00:06:02,160 feature missed the mark, so it's no longer available. 117 00:06:02,160 --> 00:06:07,200 I mean, I don't really know if AI at a base level is equal to a 118 00:06:07,200 --> 00:06:09,480 quote UN quote useful creative tool. 119 00:06:09,680 --> 00:06:12,280 And I also don't think it's a valid reason for putting out a 120 00:06:12,280 --> 00:06:14,640 feature like this. It really just goes to show how 121 00:06:14,640 --> 00:06:17,960 out of touch these tech companies are as the sag after a 122 00:06:18,000 --> 00:06:21,120 statement commented on because I feel like they're all suffering 123 00:06:21,120 --> 00:06:24,280 from groupthink at this point. Assuming that everyone wants AI 124 00:06:24,280 --> 00:06:26,520 everything, which couldn't be further from the truth. 125 00:06:26,640 --> 00:06:30,640 You can still upload images of someone and ask Meta AI to 126 00:06:30,640 --> 00:06:34,320 generate an image of them, which is obviously similar to other AI 127 00:06:34,320 --> 00:06:36,640 chat bots. But this is also something that 128 00:06:36,640 --> 00:06:39,680 Meta told me in a chat when I was trying to test this out 129 00:06:39,680 --> 00:06:41,760 while I was researching for this story. 130 00:06:41,800 --> 00:06:45,600 As the chat log read quote, you can just upload any photo you 131 00:06:45,600 --> 00:06:49,640 have and say use this person slash make this person do X and 132 00:06:49,640 --> 00:06:52,720 I'll generate from that. So while you can't tag them, I 133 00:06:52,720 --> 00:06:57,760 guess you still can be making unconsensual AI generated images 134 00:06:57,800 --> 00:07:00,280 of someone. Nevertheless, some social media 135 00:07:00,280 --> 00:07:03,440 users are touting Meta's decision to pull the feature as 136 00:07:03,440 --> 00:07:06,840 a win, as at Kate McGill vocals said on threads quote. 137 00:07:07,040 --> 00:07:10,280 It proves fighting back actually works and I definitely think 138 00:07:10,280 --> 00:07:14,840 there is some truth to that. But we can't be too naive 139 00:07:14,960 --> 00:07:17,680 because at the same time, we can't forget that this comes 140 00:07:17,920 --> 00:07:21,720 only like 2 weeks after the launch of Meta glasses, which is 141 00:07:21,720 --> 00:07:24,200 another tech product that completely undermines our 142 00:07:24,200 --> 00:07:28,200 ability to control how our likeness is used and is honestly 143 00:07:28,200 --> 00:07:30,680 just another Ave. for harassment, especially for 144 00:07:30,680 --> 00:07:33,000 women. I spoke about that in mini tech, 145 00:07:33,000 --> 00:07:36,720 I think from 2 weeks ago. So while Meta admitted that they 146 00:07:36,720 --> 00:07:39,080 quote UN quote missed the mark, it doesn't seem like this is 147 00:07:39,080 --> 00:07:42,680 something that they genuinely believe or care about. 148 00:07:43,200 --> 00:07:46,040 Considering they continue to promote products like Meta 149 00:07:46,040 --> 00:07:49,560 Glasses, it's clear where their priority lies, and it's not 150 00:07:49,600 --> 00:07:52,920 always with user safety. OK, let's get into the next 151 00:07:52,920 --> 00:07:56,560 story. In more grim news, a new report 152 00:07:56,560 --> 00:07:59,600 reveals that mainstream social media platforms are leading 153 00:07:59,600 --> 00:08:03,880 users to tools that generate non consensual intimate images, or 154 00:08:03,880 --> 00:08:08,280 in other words, nudify apps. We have had many conversations 155 00:08:08,280 --> 00:08:11,320 around platforms and features that generate not safer work or 156 00:08:11,320 --> 00:08:15,440 explicit content of social media users without their consent, and 157 00:08:15,440 --> 00:08:18,640 this did really pick up at the beginning of this year with X's 158 00:08:18,640 --> 00:08:22,360 whole undressing scandal. I did cover this whole thing in 159 00:08:22,360 --> 00:08:25,120 multiple episodes of Mini Tech Scroll, but as a quick 160 00:08:25,120 --> 00:08:28,800 refresher, similar to Muse AI, X rolled out a feature that 161 00:08:28,800 --> 00:08:33,120 allowed users to edit a photo, any photo, without permission 162 00:08:33,120 --> 00:08:35,880 from its owner or say a person in the photo. 163 00:08:35,880 --> 00:08:39,280 Unfortunately, this feature was quickly used for nefarious and 164 00:08:39,280 --> 00:08:42,360 disgusting reasons. Users started asking Grok to 165 00:08:42,360 --> 00:08:46,000 remove clothing from people featured in various images, and 166 00:08:46,000 --> 00:08:49,640 this was absolutely anyone ranging from women to minors. 167 00:08:49,640 --> 00:08:53,680 Now X kind of operates in this weird space in the online world, 168 00:08:53,680 --> 00:08:57,120 I would say these days. It obviously is increasingly 169 00:08:57,120 --> 00:09:00,680 right wing, but you do have niches like Stan Twitter that 170 00:09:00,680 --> 00:09:03,680 are still alive and well. In that sense, it is definitely 171 00:09:03,680 --> 00:09:07,560 a more mainstream platform. We're allowing its associated AI 172 00:09:07,560 --> 00:09:10,520 tool to do something like this is worrying because it's 173 00:09:10,560 --> 00:09:14,600 obviously not some random website or platform on the dark 174 00:09:14,600 --> 00:09:16,880 web or something. But this is actually part of a 175 00:09:16,880 --> 00:09:21,440 bigger trend because X isn't the only pathway to this kind of non 176 00:09:21,440 --> 00:09:24,280 consensual intimate image generation tool. 177 00:09:24,400 --> 00:09:28,280 A new report titled Mapping the New Defied Tools Ecosystem 178 00:09:28,280 --> 00:09:30,320 explored this phenomenon further. 179 00:09:30,320 --> 00:09:33,200 So for a bit of context, the report was published on Monday 180 00:09:33,200 --> 00:09:35,920 by the Institute of Strategic Dialogue, which is an 181 00:09:35,920 --> 00:09:38,880 independent research organization based in London. 182 00:09:38,880 --> 00:09:41,840 And the research they do is typically around individual 183 00:09:41,840 --> 00:09:45,560 rights, democratic process, national security and public 184 00:09:45,560 --> 00:09:47,720 safety. In this report, researchers 185 00:09:47,720 --> 00:09:52,800 explored the pathways users take to find nudified tools, which is 186 00:09:52,800 --> 00:09:56,720 obviously apps platforms features that allow users to 187 00:09:56,720 --> 00:09:59,320 generate non consensual intimate images. 188 00:09:59,680 --> 00:10:02,880 They also explored how the ecosystem operates more broadly, 189 00:10:02,880 --> 00:10:05,280 as well as how this content is circulated. 190 00:10:05,360 --> 00:10:09,360 One of the more concerning, yet perhaps unsurprising findings is 191 00:10:09,360 --> 00:10:12,640 that a lot of this takes place on mainstream social media 192 00:10:12,640 --> 00:10:16,840 platforms as opposed to more fringe platforms like 4 Chan. 193 00:10:17,160 --> 00:10:20,400 So I thought I would recap some of the key findings. 194 00:10:20,480 --> 00:10:25,040 After identifying a total of 181 websites offering tools that 195 00:10:25,040 --> 00:10:29,200 allow users to generate nudified images or videos of real people, 196 00:10:29,640 --> 00:10:33,000 the researchers found that these sites hit approximately 40 197 00:10:33,000 --> 00:10:35,320 million unique visitors per month. 198 00:10:35,480 --> 00:10:38,640 And they also found that only two of them included measures to 199 00:10:38,640 --> 00:10:41,960 verify that users were over 18 years of age. 200 00:10:42,080 --> 00:10:46,520 While it's not a majority, more than 5.7 million visits to these 201 00:10:46,520 --> 00:10:50,160 sites and apps containing notified tools came from 202 00:10:50,160 --> 00:10:52,160 mainstream social media platforms. 203 00:10:52,480 --> 00:10:56,120 YouTube X, Reddit, and Facebook are the top four, with both 204 00:10:56,120 --> 00:10:59,560 YouTube and X facilitating over 1,000,000 visits each. 205 00:10:59,960 --> 00:11:03,640 I'm curious if you guys are surprised about these platforms 206 00:11:03,640 --> 00:11:06,360 being the top four. At first glance, I was a little 207 00:11:06,360 --> 00:11:09,160 bit surprised by Facebook, but then the more I thought about it 208 00:11:09,160 --> 00:11:11,720 and considering everything that just happened with Muse AI that 209 00:11:11,720 --> 00:11:14,240 I just covered, it's actually not that surprising. 210 00:11:14,240 --> 00:11:18,440 And also, I do feel like I could see Facebook's user base going 211 00:11:18,440 --> 00:11:22,600 to platforms where you can adjust an image or generate an 212 00:11:22,600 --> 00:11:26,240 image using AI. Nevertheless, YouTube was the 213 00:11:26,240 --> 00:11:30,000 highest on this list and as the report reads, quote the top 214 00:11:30,000 --> 00:11:33,600 platform identified in terms of referrals to the top ten 215 00:11:33,600 --> 00:11:37,240 websites analyzed with YouTube, with links to the site on the 216 00:11:37,240 --> 00:11:40,600 platform resulting in over 1.82 million visits. 217 00:11:40,600 --> 00:11:43,560 Manual searches on YouTube identified numerous videos 218 00:11:43,560 --> 00:11:47,560 promoting specific Notify apps and even providing information 219 00:11:47,560 --> 00:11:50,560 on how to get Quote Undress AI promo codes. 220 00:11:51,080 --> 00:11:54,240 This seems in direct conflict with Youtubes policy restricting 221 00:11:54,240 --> 00:11:57,560 the promotion or linking of harmful external content. 222 00:11:57,640 --> 00:12:01,240 And as far as how these websites operate, a lot of them use a 223 00:12:01,400 --> 00:12:05,040 freemium model which is a token based or tiered subscription 224 00:12:05,040 --> 00:12:08,520 monetization model. So people also would have to pay 225 00:12:08,960 --> 00:12:12,360 to use these tools, whether that be through PayPal, Stripe, 226 00:12:12,520 --> 00:12:15,520 crypto, etcetera. The researchers even identified 227 00:12:15,520 --> 00:12:19,560 another layout to this monetization strategy quote in 228 00:12:19,560 --> 00:12:23,200 which individual non consensual intimate image content creators 229 00:12:23,200 --> 00:12:26,840 or resellers use third party creator economy platforms such 230 00:12:26,840 --> 00:12:31,080 as Patreon to sell access to deepfake content or generation 231 00:12:31,080 --> 00:12:33,160 services. Which is obviously very 232 00:12:33,160 --> 00:12:35,680 disturbing that they are making a business out of this. 233 00:12:35,680 --> 00:12:39,640 While YouTube had the most referral links in the data set 234 00:12:39,640 --> 00:12:42,760 for the report, X was the most prevalent social media platform 235 00:12:42,760 --> 00:12:46,440 when it came to the distribution of gendered deepfake related 236 00:12:46,440 --> 00:12:49,840 posts as well. As they explained, X was quote 237 00:12:50,080 --> 00:12:53,920 used for direct content creation request via in platform croc 238 00:12:53,920 --> 00:12:57,200 function and for promotion of notification sites. 239 00:12:57,200 --> 00:13:01,200 I presume Crocs undressing scandal is a major reason behind 240 00:13:01,200 --> 00:13:04,680 such a high number of direct content creation requests. 241 00:13:04,720 --> 00:13:08,240 At first the feature was available to everyone, then it 242 00:13:08,240 --> 00:13:10,400 was restricted, but not entirely. 243 00:13:10,680 --> 00:13:15,920 It was first limited to premium paid users, which was disturbing 244 00:13:15,920 --> 00:13:18,920 for many reasons because Gruck was still able to create the 245 00:13:18,920 --> 00:13:21,960 images and then X was essentially profiting off it. 246 00:13:21,960 --> 00:13:25,400 But since then it seemed to have been restricted further. 247 00:13:25,680 --> 00:13:30,000 X has also faced multiple lawsuits over the generation of 248 00:13:30,000 --> 00:13:32,680 these sexual non consensual images. 249 00:13:32,720 --> 00:13:36,400 And while X has said that more restrictions are in place, it 250 00:13:36,400 --> 00:13:39,280 does seem that users continue to bypass them. 251 00:13:39,600 --> 00:13:42,880 According to an investigation by Wired from just last month. 252 00:13:42,880 --> 00:13:46,360 The publication found quote dozens of nudified deepfake 253 00:13:46,360 --> 00:13:49,560 images and videos on Grok's website, including non 254 00:13:49,560 --> 00:13:52,480 consensual depictions of celebrity and at least one 255 00:13:52,480 --> 00:13:55,680 prominent US politician. So it's clear that it is an 256 00:13:55,680 --> 00:13:59,840 ongoing issue on X and I'm not sure if it will ever be fully 257 00:13:59,840 --> 00:14:03,760 resolved at this stage or at least as with X as we currently 258 00:14:03,760 --> 00:14:05,680 know it. But actually quick side note on 259 00:14:05,680 --> 00:14:07,440 X. The platform is also making 260 00:14:07,440 --> 00:14:10,600 headlines this week because they did I recently announced A tweak 261 00:14:10,600 --> 00:14:15,040 to its algorithm to boost the visibility of your mutuals, and 262 00:14:15,040 --> 00:14:16,920 many users have taken this as a win. 263 00:14:16,920 --> 00:14:20,160 But anyway, it is obviously really important to call 264 00:14:20,160 --> 00:14:23,840 attention to reports like this and findings like this because 265 00:14:24,040 --> 00:14:27,640 it does continue to be an ongoing form of harassment that 266 00:14:27,640 --> 00:14:30,080 disproportionately impacts and harms women. 267 00:14:30,360 --> 00:14:32,960 And I will also link the report in the description if you're 268 00:14:32,960 --> 00:14:34,560 interested in reading more about it. 269 00:14:34,720 --> 00:14:37,800 OK that brings me to my last story for today which is 270 00:14:37,800 --> 00:14:40,800 character AI is dabbling in micro dramas. 271 00:14:41,160 --> 00:14:43,880 We have covered micro dramas here on mini tech but I don't 272 00:14:43,880 --> 00:14:46,160 actually think we've covered character AI. 273 00:14:46,280 --> 00:14:50,080 Micro dramas are basically like mini soap operas which are 274 00:14:50,080 --> 00:14:53,080 essentially short form video clips that are part of a bigger 275 00:14:53,080 --> 00:14:55,120 series. The AI generated fruit. 276 00:14:55,120 --> 00:15:00,000 Love Island would probably be a UTC version I guess of this but 277 00:15:00,000 --> 00:15:03,000 it honestly spans way more content than that. 278 00:15:03,160 --> 00:15:06,160 For a little bit more context, it has been gaining popularity 279 00:15:06,160 --> 00:15:09,920 in China for years and only now is making its way into other 280 00:15:09,920 --> 00:15:12,120 markets. And I've also seen claims that 281 00:15:12,120 --> 00:15:15,560 micro dramas or micro series like these are the future of 282 00:15:15,560 --> 00:15:17,200 streaming. And I'd definitely be curious to 283 00:15:17,200 --> 00:15:19,600 hear what you guys think about that industry wise. 284 00:15:19,600 --> 00:15:24,000 I've also seen actors note how this kind of work is much easier 285 00:15:24,000 --> 00:15:26,800 to kind of get your foot in the door because there are so many 286 00:15:26,800 --> 00:15:29,800 more opportunities. But at the same time, I did 287 00:15:29,800 --> 00:15:32,480 actually see a TikTok earlier this week from an actor saying 288 00:15:32,480 --> 00:15:36,240 that she feels like AI micro dramas becoming increasingly 289 00:15:36,240 --> 00:15:39,600 popular because she's singing less and less, costing calls for 290 00:15:39,600 --> 00:15:42,040 these types of roles. I will inside that TikTok here 291 00:15:42,040 --> 00:15:43,880 if I can find it, you know? I was like, yeah, I'm not 292 00:15:43,880 --> 00:15:45,120 worried about AI taking our jobs. 293 00:15:45,120 --> 00:15:47,320 They're taking them. They're taking our jobs because 294 00:15:47,600 --> 00:15:51,000 there used to be 15 casting calls a day for vertical micro 295 00:15:51,000 --> 00:15:55,120 dramas, whatever. Now there's one, maybe 2, and 296 00:15:55,120 --> 00:15:58,000 that's my main form of income. So I'm a little bit. 297 00:15:58,000 --> 00:16:00,840 Nervous, I can't lie. It's been it's been fast, but 298 00:16:00,840 --> 00:16:03,920 also slow. Like I'm getting auditions for 299 00:16:03,920 --> 00:16:08,000 bigger stuff, which is exciting, except for that isn't not 300 00:16:08,000 --> 00:16:11,280 consistent income. Like yes, a big movie audition 301 00:16:11,280 --> 00:16:15,360 is like so exciting. The likelihood of booking that. 302 00:16:15,800 --> 00:16:18,720 Or like. That paying my next month's rent 303 00:16:19,160 --> 00:16:22,480 low, so that's a little bit scary. 304 00:16:23,040 --> 00:16:26,640 Now Character AI is an AI platform where you can build, 305 00:16:26,840 --> 00:16:29,720 customize, and chat with different characters, and they 306 00:16:29,720 --> 00:16:32,920 may be original characters, they could be historical figures or 307 00:16:32,920 --> 00:16:36,040 anything in between. I remember Brittany Broski made 308 00:16:36,040 --> 00:16:38,840 a couple videos using Character AIA few years ago. 309 00:16:38,840 --> 00:16:41,040 That was really when it came under my radar. 310 00:16:41,160 --> 00:16:43,360 Personally. I remember she even made a 311 00:16:43,360 --> 00:16:46,840 character on Character AI but I can't remember what it was 312 00:16:46,840 --> 00:16:50,360 called. If anyone remembers please let 313 00:16:50,360 --> 00:16:53,200 me know. But Character AI has faced 314 00:16:53,200 --> 00:16:55,200 significant controversy since then. 315 00:16:55,320 --> 00:16:59,640 In 2024, a mother filed a civil suit against Character AI, 316 00:16:59,640 --> 00:17:03,080 alleging that her 14 year old son died by suicide after 317 00:17:03,080 --> 00:17:06,200 communicating with the chatbot. The family has since settled the 318 00:17:06,200 --> 00:17:09,560 lawsuit, and Character AI has also announced new safety 319 00:17:09,560 --> 00:17:12,280 features over the years. And now it seems like they're 320 00:17:12,280 --> 00:17:15,839 pivoting a little bit and trying their hand at micro dramas. 321 00:17:15,839 --> 00:17:20,160 So last week Character AI announced its CAI series which 322 00:17:20,160 --> 00:17:24,160 is quote a new micro dramas offering of original short form 323 00:17:24,160 --> 00:17:27,319 vertical series created by our in house studio team. 324 00:17:27,440 --> 00:17:31,120 Per the platform's blog post, they also add that the team is 325 00:17:31,120 --> 00:17:34,360 made-up of people with credits across Netflix, Nickelodeon, 326 00:17:34,360 --> 00:17:37,240 DreamWorks and more. Though I should also say that 327 00:17:37,240 --> 00:17:40,160 Character AI does note that these dramas were also created 328 00:17:40,160 --> 00:17:43,000 with AI production tools, so it does seem to be a mixture of 329 00:17:43,000 --> 00:17:46,200 both human and AI. To start up this offering, there 330 00:17:46,200 --> 00:17:49,360 are three shows. The 1st is called last summer 331 00:17:49,360 --> 00:17:52,480 which is a romance anime, the second one is called the 332 00:17:52,480 --> 00:17:55,600 nighttime game which is a quote UN quote Gen. 333 00:17:55,600 --> 00:18:00,240 Z paranormal horror and the last one is Eden Fall which is a 334 00:18:00,360 --> 00:18:04,080 survival drama on why character AI is kind of going down this 335 00:18:04,080 --> 00:18:06,080 path. The Hollywood Reporter 336 00:18:06,080 --> 00:18:09,200 hypothesizes that quote. The move into micro drama seems 337 00:18:09,200 --> 00:18:11,840 like an attempt to diversify a business that is facing 338 00:18:11,840 --> 00:18:15,240 significant headwinds, moving it beyond the kinds of intimate 339 00:18:15,240 --> 00:18:19,080 companion style relationships that give rise to many problems. 340 00:18:19,120 --> 00:18:21,640 And while that might be part of the reasoning, I should note 341 00:18:21,640 --> 00:18:24,520 that they aren't just producing these micro dramas, they are 342 00:18:24,520 --> 00:18:26,160 doing things a little bit differently. 343 00:18:26,280 --> 00:18:29,960 After watching over 18, users can chat with characters from 344 00:18:30,000 --> 00:18:33,760 each series on character AI, and it seems to be that they can 345 00:18:33,760 --> 00:18:36,320 really talk about anything, whether that be the characters 346 00:18:36,320 --> 00:18:39,160 relationship with other characters, talk about specific 347 00:18:39,160 --> 00:18:44,040 moments, or even role play storylines from the series. 348 00:18:44,040 --> 00:18:47,160 It's kind of giving interactive fan fiction in that respect. 349 00:18:47,240 --> 00:18:50,640 As the blog post continues, quote that creates A deeper 350 00:18:50,640 --> 00:18:53,560 connection between fans and characters and gives our studio 351 00:18:53,560 --> 00:18:56,960 team a new feedback loop as we learn which characters, worlds, 352 00:18:56,960 --> 00:18:59,120 and stories people want to spend more time with. 353 00:18:59,280 --> 00:19:03,320 I will say this does feel very 2026, but better off or worse 354 00:19:03,320 --> 00:19:06,640 with the crossover of AI chat bots and micro dramas. 355 00:19:06,720 --> 00:19:09,760 I do think there is quite a slippery slope here because 356 00:19:09,840 --> 00:19:13,720 character AI has already been through a wrongful death lawsuit 357 00:19:13,720 --> 00:19:18,120 as I've covered, which is tied to the parasocial relationship 358 00:19:18,280 --> 00:19:20,520 ship someone develops with a chat bot. 359 00:19:20,520 --> 00:19:23,920 I don't even know if parasocial is the right word here, but it 360 00:19:23,920 --> 00:19:26,480 is some kind of deep emotional attachment. 361 00:19:26,480 --> 00:19:29,280 And this new format and the ability to chat with the 362 00:19:29,280 --> 00:19:33,400 characters after engaging with the series is arguably built on 363 00:19:33,400 --> 00:19:36,520 that same dynamic and engagement strategy. 364 00:19:36,600 --> 00:19:40,000 For a platform that maybe should try to limit these relationships 365 00:19:40,000 --> 00:19:42,000 a little bit. That format doesn't really seem 366 00:19:42,000 --> 00:19:46,240 to do that at first glance. I'd be curious to see how people 367 00:19:46,240 --> 00:19:49,560 interact with it if they do. I also do think this 368 00:19:49,560 --> 00:19:52,080 conversation is related to a bigger conversation we're 369 00:19:52,080 --> 00:19:55,600 constantly having around AI companions, like AI girlfriends 370 00:19:55,600 --> 00:19:59,040 that a lot of young men are building unhealthy relationships 371 00:19:59,040 --> 00:20:00,360 with. But I'm definitely curious to 372 00:20:00,360 --> 00:20:03,360 hear what you guys think about this and the connection of micro 373 00:20:03,360 --> 00:20:05,720 dramas and talking with those characters. 374 00:20:05,720 --> 00:20:08,480 Whether you think that is something that we could see down 375 00:20:08,480 --> 00:20:12,120 the line with more mainstream TV shows or is that just a very 376 00:20:12,120 --> 00:20:14,880 like chronically online thing? OK, that brings us. 377 00:20:14,960 --> 00:20:17,480 To the end of Mini Tech Scroll for today. 378 00:20:17,480 --> 00:20:20,560 I really hope you guys enjoyed. Lauren will speak to you 379 00:20:20,560 --> 00:20:22,680 tomorrow and I will be back next week. 380 00:20:22,880 --> 00:20:24,400 Thanks for listening. Bye. 381 00:20:24,400 --> 00:20:27,400 Centennial World Podcast Network operates on Gadigal land. 382 00:20:27,680 --> 00:20:30,680 We would like to acknowledge the Gadigal of the Aura Nation, the 383 00:20:30,680 --> 00:20:33,840 traditional custodians of this land, and pay our respects to 384 00:20:33,840 --> 00:20:35,640 elders both past and present.