1 00:00:09,119 --> 00:00:11,940 S1: Hey there. I'm Osman Farooqui and this is The Drop, 2 00:00:11,940 --> 00:00:14,400 S1: a weekly culture show from the Sydney Morning Herald in 3 00:00:14,400 --> 00:00:17,070 S1: The Age, where we dive into the latest in the 4 00:00:17,070 --> 00:00:21,600 S1: world of pop culture and entertainment, ie it's all the 5 00:00:21,600 --> 00:00:25,140 S1: rage at the moment, from business to art to journalism 6 00:00:25,140 --> 00:00:29,160 S1: and even TikTok. Artificial intelligence is being used to create, 7 00:00:29,160 --> 00:00:34,500 S1: sometimes hilarious and sometimes downright terrifying content. But is it 8 00:00:34,500 --> 00:00:37,290 S1: something to actually fear, or is it another tool that 9 00:00:37,290 --> 00:00:40,140 S1: we can embrace, particularly in the world of arts and culture, 10 00:00:40,140 --> 00:00:42,720 S1: where it can be really helpful for all sorts of 11 00:00:42,720 --> 00:00:46,140 S1: different creators? To help figure that out, I'm joined this 12 00:00:46,140 --> 00:00:48,870 S1: week by my colleague Jim Marlow, a reporter at The 13 00:00:48,870 --> 00:00:52,260 S1: Age and the Sydney Morning Herald. Jim Marlow, how are 14 00:00:52,260 --> 00:00:53,550 S1: you going? Welcome to the Drop. 15 00:00:53,550 --> 00:00:55,350 S2: I'm going great. Osman Thank. 16 00:00:55,350 --> 00:00:57,780 S1: You. Good to have you on the show. So last 17 00:00:57,780 --> 00:01:00,990 S1: week you alerted me to this TikTok trend that I 18 00:01:00,990 --> 00:01:04,530 S1: think is actually a really interesting kind of confluence of 19 00:01:04,530 --> 00:01:08,160 S1: these different threads that are playing out across social media 20 00:01:08,160 --> 00:01:12,300 S1: and in technology and in the arts and the trend, basically, 21 00:01:12,300 --> 00:01:13,920 S1: it's going to sound a bit weird, but it basically 22 00:01:13,920 --> 00:01:20,190 S1: involves AI generated voices of different US presidents being kind 23 00:01:20,190 --> 00:01:24,620 S1: of toxic to each other in a video game chat room, 24 00:01:24,620 --> 00:01:26,760 S1: a discord type thing. Is that basically. 25 00:01:26,760 --> 00:01:29,700 S2: Right? Yep, 100%. It's this really fascinating, sort of like 26 00:01:29,850 --> 00:01:33,360 S2: look at how AI voice generation is being used currently. 27 00:01:33,360 --> 00:01:34,560 S2: I mean, at the moment it's a bit of like 28 00:01:34,560 --> 00:01:37,590 S2: a parlor trick, right? It's not really doesn't seem serious 29 00:01:37,590 --> 00:01:40,830 S2: and scary yet, but I mean, yeah, that could change. 30 00:01:45,200 --> 00:01:46,880 S3: Joe, you know I love you, but that beat was 31 00:01:46,880 --> 00:01:47,580 S3: a bit mid. 32 00:01:47,600 --> 00:01:50,300 S4: That beat was so ass. You were definitely asleep while 33 00:01:50,300 --> 00:01:51,140 S4: you made that. 34 00:01:51,260 --> 00:01:53,330 S3: Shut the fuck up, Donald. What do you even know 35 00:01:53,330 --> 00:01:54,770 S3: about making sick beats? 36 00:01:54,860 --> 00:01:57,650 S4: The 808 was out of key. The melody sounds like 37 00:01:57,650 --> 00:02:01,610 S4: something a newborn baby would create accidentally. That clap sample 38 00:02:01,610 --> 00:02:02,390 S4: makes me want to kill. 39 00:02:02,540 --> 00:02:06,800 S1: What is it about these particular. Clips of these videos. 40 00:02:06,800 --> 00:02:09,530 S1: It's sort of it's become its own universe in a way. 41 00:02:09,530 --> 00:02:12,500 S1: These things are kind of ongoing. What? I mean, getting 42 00:02:12,500 --> 00:02:15,020 S1: someone to explain how a joke is funny is never 43 00:02:15,020 --> 00:02:18,260 S1: really a fruitful task. But there's something about this in 44 00:02:18,260 --> 00:02:21,350 S1: the way that is it, the people involved, the fact 45 00:02:21,350 --> 00:02:23,630 S1: that the AI is weirdly quite good, the topics that 46 00:02:23,630 --> 00:02:25,730 S1: they're talking about, all of the above that make this 47 00:02:25,730 --> 00:02:29,059 S1: both so funny to the extent that it's got millions 48 00:02:29,060 --> 00:02:30,079 S1: and millions of views. 49 00:02:30,080 --> 00:02:32,030 S2: Yeah, I would say that's probably the case. I mean, 50 00:02:32,030 --> 00:02:34,640 S2: like you're right that explaining the joke kind of ruins 51 00:02:34,639 --> 00:02:36,050 S2: it and I feel like I've been ruining it for 52 00:02:36,050 --> 00:02:38,030 S2: myself a little bit in the past the past couple 53 00:02:38,030 --> 00:02:41,630 S2: of days thinking about this conversation. And yeah, it does 54 00:02:41,630 --> 00:02:43,250 S2: sort of boil down to the fact that it's like 55 00:02:43,250 --> 00:02:46,549 S2: just kind of funny to hear like us presidents argue 56 00:02:46,550 --> 00:02:48,530 S2: with each other about just, you know, random stuff that 57 00:02:48,530 --> 00:02:51,260 S2: you might I'm not sure if you're into gaming, but, 58 00:02:51,620 --> 00:02:53,810 S2: you know, if you're like in a voice chat, like 59 00:02:53,810 --> 00:02:56,210 S2: you're always like needling each other and having a go. 60 00:02:56,210 --> 00:02:58,970 S2: And it's funny to see these dynamics play out with 61 00:02:58,970 --> 00:03:02,750 S2: the president, says the actors. And I just think it's 62 00:03:02,750 --> 00:03:05,000 S2: generally funny if you are like the member of like 63 00:03:05,090 --> 00:03:09,169 S2: a subculture to sort of take like an established character 64 00:03:09,169 --> 00:03:12,290 S2: and then sort of insert them into your subculture and 65 00:03:12,290 --> 00:03:15,110 S2: how they might react and and what positions they might take. 66 00:03:15,110 --> 00:03:17,810 S2: So for example, one of the one of the tiktoks 67 00:03:17,810 --> 00:03:22,100 S2: I showed you was the president's playing Fortnite and talking 68 00:03:22,100 --> 00:03:25,340 S2: about who is the the best like rapper, right as 69 00:03:25,340 --> 00:03:27,500 S2: a whole, or is it Kendrick Lamar? 70 00:03:27,800 --> 00:03:29,780 S5: J. Cole doesn't have a single piece of work that 71 00:03:29,780 --> 00:03:32,390 S5: can go toe to toe with Tab or good kid 72 00:03:32,389 --> 00:03:34,130 S5: Mad City. Stay crying. 73 00:03:34,150 --> 00:03:37,580 S3: Sleep to the chocolate factory, you delusional Oompa Loompa. Cole 74 00:03:37,580 --> 00:03:39,920 S3: is hip hop. Oh yeah. Kendrick puts on a dumb 75 00:03:39,920 --> 00:03:41,840 S3: ass voice for a song and you call him a. 76 00:03:41,840 --> 00:03:44,030 S5: Genius for it. And how many Pulitzer Prizes does Jay 77 00:03:44,030 --> 00:03:46,070 S5: flop have out of here? Yeah, that's what I think. 78 00:03:46,070 --> 00:03:48,590 S5: How about Macklemore? Guys, Macklemore is so cool. 79 00:03:48,590 --> 00:03:51,020 S6: I love Thrift Shop. I got it on my iPod. 80 00:03:51,410 --> 00:03:53,620 S2: Here and just like it hit a number of like, 81 00:03:53,630 --> 00:03:56,180 S2: pleasure centers for me, but like, it just sort of 82 00:03:56,270 --> 00:03:57,770 S2: like stuck with me because it was like it was 83 00:03:57,770 --> 00:04:01,550 S2: interesting to see like, who they characterize as having what 84 00:04:01,550 --> 00:04:05,300 S2: opinion and how they sort of like all played off 85 00:04:05,300 --> 00:04:07,910 S2: each other. And one of the key things that I 86 00:04:08,150 --> 00:04:09,890 S2: that sort of made me want to write about it 87 00:04:09,890 --> 00:04:12,470 S2: or talk to you about it was that, you know, 88 00:04:12,470 --> 00:04:14,930 S2: it's it's kind of fascinating to see how they like, 89 00:04:14,930 --> 00:04:17,180 S2: choose to characterize someone like Trump, who is like a 90 00:04:17,180 --> 00:04:17,750 S2: very like. 91 00:04:17,750 --> 00:04:19,340 S1: So who's his favorite rapper in these videos? 92 00:04:19,370 --> 00:04:22,310 S2: Yeah he's he's Kendrick fan obviously like he calls him 93 00:04:22,400 --> 00:04:24,980 S2: Yeah yeah And it's funny because like they position Trump 94 00:04:24,980 --> 00:04:28,070 S2: as like being sort of like morally and like depending 95 00:04:28,070 --> 00:04:30,650 S2: on who you ask, like, objectively correct in this sense. Yeah. 96 00:04:30,800 --> 00:04:33,890 S2: And then like those, like, I don't know, it's the 97 00:04:33,890 --> 00:04:36,410 S2: thing that, like, I'm really interested in is this the 98 00:04:36,410 --> 00:04:38,570 S2: way they're sort of like whitewashing Trump's image, right? Like, 99 00:04:38,570 --> 00:04:40,100 S2: I know I'm getting a little bit off course here, 100 00:04:40,100 --> 00:04:40,970 S2: but like I'm just sort of. 101 00:04:41,540 --> 00:04:43,310 S1: There's an extra layer to it. It's like, yeah, it's 102 00:04:43,310 --> 00:04:45,380 S1: like you're like, Oh, yeah, he's right. This is a 103 00:04:45,380 --> 00:04:48,890 S1: guy arguing really coherently about why Kendrick is the greatest 104 00:04:49,040 --> 00:04:51,290 S1: of all time. But you're like, Hang on, that's Donald 105 00:04:51,320 --> 00:04:53,150 S1: Trump's voice. This is really odd. 106 00:04:53,150 --> 00:04:55,789 S2: Yeah, it's bizarre, isn't it? Yeah. And it's, um, I 107 00:04:55,790 --> 00:04:59,360 S2: think it speaks to his position in the wider political 108 00:04:59,360 --> 00:05:01,190 S2: landscape at this point because, you know, he is in 109 00:05:01,190 --> 00:05:04,520 S2: some ways like a like politically emasculated. Right? He doesn't 110 00:05:04,520 --> 00:05:06,710 S2: really have the power that he had, just like, what 111 00:05:06,710 --> 00:05:09,810 S2: is it, three, three, four years ago. Yeah. And, you know, 112 00:05:09,830 --> 00:05:11,960 S2: he becomes a bit of a punchline. But it's interesting 113 00:05:11,960 --> 00:05:14,029 S2: to see how that will sort of play out in 114 00:05:14,029 --> 00:05:14,359 S2: the years ahead. 115 00:05:14,390 --> 00:05:17,510 S1: Yeah. And it's not just current contemporary political figures like 116 00:05:17,690 --> 00:05:20,479 S1: Obama pops up in these videos, George Bush pops up 117 00:05:20,480 --> 00:05:20,750 S1: as well. 118 00:05:20,750 --> 00:05:23,030 S2: Obama's characterization is by far the funniest to me. 119 00:05:23,029 --> 00:05:26,060 S7: But Drake, I just wanted to say I love your music. 120 00:05:26,060 --> 00:05:28,220 S7: Michelle and I used to get turnt to hotline bling 121 00:05:28,220 --> 00:05:30,260 S7: while reigning drone strikes on hospitals. 122 00:05:30,589 --> 00:05:32,870 S2: Obviously there's quite a lot of criticism of his like 123 00:05:32,870 --> 00:05:36,979 S2: a track record when it comes to wars. And in 124 00:05:36,980 --> 00:05:39,890 S2: these tiktoks they're constantly sort of characterizing as someone who 125 00:05:39,890 --> 00:05:42,409 S2: like has a, you know, a very sort of itchy 126 00:05:42,410 --> 00:05:46,520 S2: trigger finger. He's always, you know, bombing just random people overseas. 127 00:05:46,520 --> 00:05:50,360 S2: And yeah, it's funny to say that he isn't off 128 00:05:50,360 --> 00:05:52,849 S2: the hook for his, you know, his political it's really interesting. 129 00:05:53,000 --> 00:05:56,210 S1: It feels like to really I guess this is what 130 00:05:56,210 --> 00:06:00,500 S1: I'm interested in, to really get every layer of this joke. 131 00:06:00,500 --> 00:06:03,530 S1: You have to sort of be pretty aware of current 132 00:06:03,529 --> 00:06:08,090 S1: online gaming culture, have some pretty deep thoughts on this 133 00:06:08,089 --> 00:06:11,539 S1: state of hip hop and who the goats are, and 134 00:06:11,540 --> 00:06:16,130 S1: to have a pretty specific sense of the politics of 135 00:06:16,130 --> 00:06:21,560 S1: the last 25 years of American presidents. And, you know, 136 00:06:21,560 --> 00:06:25,070 S1: sometimes with jokes that specificity is where the the real 137 00:06:25,070 --> 00:06:27,470 S1: sweet spot lies. Like he said, it activates his pleasure centers. 138 00:06:27,470 --> 00:06:29,930 S1: Because if you if you are at that Venn diagram 139 00:06:29,930 --> 00:06:31,820 S1: of loving all of those three things or being interested 140 00:06:31,820 --> 00:06:34,190 S1: in them, as I think you are, as I am, 141 00:06:34,190 --> 00:06:36,410 S1: as clearly millions of people, I it's almost like that 142 00:06:36,410 --> 00:06:38,960 S1: joke has been created for you. Yeah. 143 00:06:38,960 --> 00:06:41,300 S2: And I think that's part of the, like, the joy 144 00:06:41,300 --> 00:06:44,990 S2: in these memes. I know, like memes are like, I 145 00:06:44,990 --> 00:06:47,690 S2: may be controversial, but I do consider them to be like, 146 00:06:47,690 --> 00:06:48,500 S2: artistic in a way. 147 00:06:48,529 --> 00:06:50,779 S1: That's a controversial. No, of course, 100% wonderful. 148 00:06:50,779 --> 00:06:53,510 S2: The safe space. I also think that they can be 149 00:06:53,510 --> 00:06:56,510 S2: considered political cartoons in a way as well. Which political 150 00:06:56,510 --> 00:06:57,490 S2: cartoonist hate it when I say. 151 00:06:57,589 --> 00:06:59,120 S1: Maybe they should get better at making memes. 152 00:06:59,330 --> 00:07:01,760 S2: Yeah, well, they should pick up some AI generation tools. 153 00:07:02,510 --> 00:07:04,430 S2: They also probably would hate me saying that, too. But, 154 00:07:04,430 --> 00:07:06,370 S2: you know, I think. It sort of speaks to this ability, 155 00:07:06,370 --> 00:07:09,640 S2: like it's the democratization of art, like, you know, the 156 00:07:09,640 --> 00:07:12,400 S2: barrier to getting involved in art constantly gets lowered and 157 00:07:12,400 --> 00:07:14,770 S2: therefore more people have the opportunity to become a part 158 00:07:14,770 --> 00:07:18,130 S2: of it. You know, niche Internet communities have always been 159 00:07:18,130 --> 00:07:20,770 S2: making memes for since memes were a thing and since 160 00:07:20,770 --> 00:07:23,470 S2: we decided that's what we wanted to call them. And 161 00:07:23,470 --> 00:07:25,989 S2: I really think that the sort of the more the 162 00:07:25,990 --> 00:07:28,540 S2: bar gets lower, the more people can get involved and 163 00:07:28,540 --> 00:07:31,660 S2: the more, you know, pleasure centers, you know, the memes 164 00:07:31,660 --> 00:07:33,370 S2: will be able to hit for specific. 165 00:07:33,400 --> 00:07:35,320 S1: That's really interesting. Yeah. To, to kind of make stuff 166 00:07:35,320 --> 00:07:39,340 S1: like this five, ten, 20 years ago required a very 167 00:07:39,340 --> 00:07:41,380 S1: specific set of tools and probably a lot of money 168 00:07:41,380 --> 00:07:43,720 S1: and resources to make something like this. But when you 169 00:07:43,720 --> 00:07:47,470 S1: can use AI and TikTok video editing, the kinds of 170 00:07:47,470 --> 00:07:51,180 S1: people who make art shifts, is this basically what you're saying? 171 00:07:51,190 --> 00:07:53,710 S2: Yeah. Yeah. And which I think is fascinating, but there 172 00:07:53,710 --> 00:07:57,730 S2: is like some questions about how that undermines, like the 173 00:07:57,730 --> 00:08:01,420 S2: act of creation itself when it comes to like making art. 174 00:08:01,780 --> 00:08:04,480 S2: But it so far, I mean, like for me, scrolling 175 00:08:04,480 --> 00:08:07,900 S2: through my TikTok feed, it's, you know, unambiguous. Plus. 176 00:08:09,040 --> 00:08:11,860 S1: You're right that there are bigger themes here. Like we're 177 00:08:11,860 --> 00:08:15,940 S1: talking about how funny and and and engaging in political 178 00:08:15,940 --> 00:08:18,520 S1: and interesting these tools can be. There's obviously all sorts 179 00:08:18,520 --> 00:08:21,310 S1: of other consequences as well. But before we get into that, 180 00:08:21,310 --> 00:08:25,360 S1: I think a really interesting example of how democratized this 181 00:08:25,360 --> 00:08:28,750 S1: is and how easy it is to make art using AI. 182 00:08:28,780 --> 00:08:31,570 S1: Like you've got an example, an app that someone you 183 00:08:31,570 --> 00:08:32,470 S1: know has made. 184 00:08:32,800 --> 00:08:36,310 S2: What they've created is this, this I sort of wrapping bot. 185 00:08:36,309 --> 00:08:38,559 S2: So effectively all you have to do is put in 186 00:08:38,559 --> 00:08:40,000 S2: like a prompt of what you want the song to 187 00:08:40,000 --> 00:08:41,709 S2: be about. And you can also put in some sort 188 00:08:41,710 --> 00:08:44,380 S2: of like markers of what you want the beat to 189 00:08:44,380 --> 00:08:46,449 S2: sound like and it will spit out like a very short, 190 00:08:46,450 --> 00:08:48,130 S2: like four bar verse. 191 00:08:48,160 --> 00:08:49,360 S1: Can we, can we have a look at it? Can 192 00:08:49,360 --> 00:08:49,950 S1: we do it now? 193 00:08:49,960 --> 00:08:51,699 S2: Yeah, absolutely. So what do you want to hear? 194 00:08:51,820 --> 00:08:53,740 S1: What sort of inputs can I give to the beat? 195 00:08:54,070 --> 00:08:55,210 S2: Trap beats work. Well. 196 00:08:55,510 --> 00:08:56,890 S1: Like if it's a trap beat. Okay. 197 00:08:57,220 --> 00:08:59,500 S2: Do you want to anything in particular you want the 198 00:08:59,500 --> 00:09:00,160 S2: subject to be? 199 00:09:00,429 --> 00:09:02,620 S1: Yeah, I think I want the subject to be you. 200 00:09:02,620 --> 00:09:06,250 S1: I want it to be. Is Jim a good journalist? Oh, 201 00:09:06,250 --> 00:09:10,510 S1: my God. So you're just typing that phrase into this app? 202 00:09:10,530 --> 00:09:15,840 S2: Yes, Against my will. Is Jim good journalist with questions? 203 00:09:15,840 --> 00:09:17,010 S2: I don't think it does so well. I did a 204 00:09:17,010 --> 00:09:19,860 S2: good one the other day that was like I ate 205 00:09:19,890 --> 00:09:22,260 S2: too much chicken feet at Yamaha and like that was 206 00:09:22,260 --> 00:09:24,210 S2: actually pretty good, but I couldn't say MSG. But anyway, 207 00:09:24,210 --> 00:09:24,720 S2: here we go. 208 00:09:31,320 --> 00:09:35,160 UU: He's trying to tell the truth, but not enough. He. 209 00:09:35,520 --> 00:09:38,130 UU: An offensive. Pretty. That's what he. 210 00:09:38,130 --> 00:09:41,300 S8: Does. He's never afraid to take a stand. 211 00:09:41,309 --> 00:09:44,820 UU: No matter the cost. He's a voice of the people 212 00:09:44,820 --> 00:09:46,589 UU: and will never be lost. 213 00:09:46,920 --> 00:09:49,830 S1: Wow. I mean, Wheatley does feel like it applies to you. 214 00:09:50,340 --> 00:09:53,840 S2: Thanks, man. I appreciate it. Yeah. So this is in beta, 215 00:09:53,850 --> 00:09:55,770 S2: I just will add. So some of the words do 216 00:09:55,770 --> 00:09:58,830 S2: come out a little bit weird, but it's pretty interesting technology. 217 00:09:58,830 --> 00:09:59,360 S2: And that was quick. 218 00:09:59,370 --> 00:10:01,290 S1: You said it takes a second. That took like a second. 219 00:10:01,290 --> 00:10:04,050 S2: Yeah, pretty much. It's I would say less than a minute. 220 00:10:04,050 --> 00:10:06,780 S2: Like about 30 seconds maybe. Yeah, It's it's pretty interesting. 221 00:10:06,780 --> 00:10:09,390 S1: And I think what is interesting about that and the 222 00:10:09,390 --> 00:10:12,630 S1: TikTok videos we were talking about is like both of these, 223 00:10:12,630 --> 00:10:16,740 S1: as we were saying, are pretty irreverent examples of using 224 00:10:16,740 --> 00:10:20,040 S1: this kind of technology, using this kind of approach to 225 00:10:20,040 --> 00:10:22,890 S1: making a meme or making art, making a political statement. 226 00:10:23,160 --> 00:10:27,090 S1: But they're also part of a much bigger conversation about 227 00:10:27,090 --> 00:10:30,090 S1: where we are going with this, with AI and tech 228 00:10:30,090 --> 00:10:34,439 S1: and this blurring of fact and fiction. When you're watching 229 00:10:34,440 --> 00:10:37,920 S1: those videos, we're talking about the presidential ones. When you're 230 00:10:37,920 --> 00:10:41,160 S1: seeing people make raps like this, I mean, is there 231 00:10:41,160 --> 00:10:43,650 S1: any part of you that thinks as funny and wild 232 00:10:43,650 --> 00:10:45,870 S1: as this is, this also has the potential to be 233 00:10:45,870 --> 00:10:47,070 S1: kind of terrifying? 234 00:10:47,070 --> 00:10:51,540 S2: Yeah. Look, I, I feel at this point, it's like 235 00:10:51,540 --> 00:10:54,960 S2: it's very interesting to see, like, which areas of life 236 00:10:54,960 --> 00:10:58,500 S2: are being affected by it first, because you would think 237 00:10:58,500 --> 00:11:02,040 S2: that I would be a tool, right? And not necessarily 238 00:11:02,040 --> 00:11:04,560 S2: a sort of like a plaything. We're sort of at 239 00:11:04,559 --> 00:11:06,720 S2: a stage where it's like it sort of is seeping 240 00:11:06,720 --> 00:11:09,959 S2: into art first because it has little consequence, right? Like 241 00:11:10,410 --> 00:11:13,260 S2: what is being done with it? You know, art does 242 00:11:13,260 --> 00:11:16,319 S2: have consequence and it matters. And I would argue it matters. 243 00:11:16,320 --> 00:11:18,930 S2: And when I was going on my tangent about the 244 00:11:18,929 --> 00:11:21,700 S2: characterization of Donald Trump, right, yeah, it matters because like, 245 00:11:21,780 --> 00:11:23,610 S2: it does influence how we see the world and how 246 00:11:23,610 --> 00:11:26,490 S2: we think about things. But the fact that you can 247 00:11:26,490 --> 00:11:29,220 S2: just spit out something funny to show your mates is 248 00:11:29,340 --> 00:11:32,460 S2: or your niche Internet community, it's a fun thing you 249 00:11:32,460 --> 00:11:36,120 S2: can do and you can just sort of enjoy yourself 250 00:11:36,120 --> 00:11:36,960 S2: and make others. 251 00:11:36,960 --> 00:11:39,960 S1: Happy with it. Yeah, it's interesting that when GPT was released, 252 00:11:39,960 --> 00:11:42,870 S1: that's the kind of service where you can sort of 253 00:11:42,870 --> 00:11:45,990 S1: type in prompts and it can create essays and there's 254 00:11:45,990 --> 00:11:47,729 S1: a little bit of a panic about what that means 255 00:11:47,730 --> 00:11:50,189 S1: for academia and journalism. But everyone I know just use 256 00:11:50,190 --> 00:11:51,540 S1: that to make like gags. 257 00:11:51,540 --> 00:11:54,750 S2: One of our colleagues actually wrote an article using Chatgpt 258 00:11:54,750 --> 00:11:58,200 S2: and like, you know, it's weirdly good, but like, at 259 00:11:58,200 --> 00:12:00,240 S2: this point it really only like, puts up a facade 260 00:12:00,240 --> 00:12:03,780 S2: of intelligence. And it's just it's a mimicry of and, 261 00:12:03,780 --> 00:12:06,809 S2: you know, in some cases that's enough, right? But in 262 00:12:06,809 --> 00:12:09,300 S2: most cases it's not really developed enough to be a 263 00:12:09,300 --> 00:12:12,120 S2: serious tool. I think. I think it's a good jumping 264 00:12:12,120 --> 00:12:14,640 S2: off point and can give you the frames to build 265 00:12:14,640 --> 00:12:17,610 S2: something greater. But it's it's not really anything more serious 266 00:12:17,610 --> 00:12:19,260 S2: than that. And I think that applies to art as well. 267 00:12:19,260 --> 00:12:23,099 S2: I mean, like memes, a low effort, low reward. So 268 00:12:23,100 --> 00:12:25,500 S2: it doesn't matter that, you know, our presents sound a 269 00:12:25,500 --> 00:12:27,630 S2: bit weird or, you know, the rapper made in 30 270 00:12:27,630 --> 00:12:30,360 S2: seconds sounds a little bit weird. He couldn't pronounce integrity, right? Like, 271 00:12:30,750 --> 00:12:32,850 S2: that's not the end of the world. But I feel 272 00:12:32,850 --> 00:12:34,439 S2: like we're still a few steps before we have to 273 00:12:34,440 --> 00:12:36,540 S2: start worrying about it, you know, taking our jobs or 274 00:12:36,660 --> 00:12:40,560 S2: I think the concern is propaganda, right? Like can it 275 00:12:40,559 --> 00:12:42,630 S2: lowers the bar for people to create propaganda. 276 00:12:42,660 --> 00:12:44,820 S1: That I think is the most interesting. Like the chat 277 00:12:44,820 --> 00:12:48,960 S1: GPT thing hasn't really worried me. Like academic seem slightly 278 00:12:48,960 --> 00:12:51,359 S1: stressed about kids faking it for essays. I mean, like 279 00:12:51,360 --> 00:12:54,720 S1: kids have found ways to cheat on essays forever. Honestly, 280 00:12:54,720 --> 00:12:57,900 S1: who goes and seriously, who cares? University is a joke. 281 00:12:57,900 --> 00:12:59,700 S1: Sorry if you are studying when you listen to this. 282 00:13:00,960 --> 00:13:02,970 S1: It's weird. It's funny that like, the first thing that's 283 00:13:02,970 --> 00:13:05,550 S1: freaked me out about all of this stuff is these 284 00:13:05,550 --> 00:13:09,240 S1: TikTok videos you've said about Donald Trump rapping Donald Trump, 285 00:13:09,240 --> 00:13:13,710 S1: talking about his favorite rappers. Yeah, because it's so realistic. 286 00:13:13,830 --> 00:13:15,780 S1: I guess, like it's two different issues here. It's what's 287 00:13:15,780 --> 00:13:18,000 S1: separating these things out because we're kind of alluding to 288 00:13:18,000 --> 00:13:19,560 S1: both of them, but like, let's tackle them one at 289 00:13:19,559 --> 00:13:21,569 S1: a time. The first one is kind of this use 290 00:13:21,570 --> 00:13:25,920 S1: of deepfakes, which obviously has the ability to create humorous 291 00:13:25,920 --> 00:13:29,370 S1: videos but also could have political and social consequences. And 292 00:13:29,370 --> 00:13:31,980 S1: then there's a separate issue of art and what it 293 00:13:31,980 --> 00:13:35,220 S1: means to have tools and artificial intelligence that can generate 294 00:13:35,220 --> 00:13:37,710 S1: a lot of art very, very quickly. Let's talk about 295 00:13:37,710 --> 00:13:42,690 S1: deepfakes first. Like, aside from the fun stuff we've been discussing. 296 00:13:43,290 --> 00:13:46,260 S1: There's all sorts of examples that have already gone viral 297 00:13:46,260 --> 00:13:49,620 S1: from kind of more nefarious or weirder ones like this. 298 00:13:49,620 --> 00:13:52,650 S1: This one I remember Joe Rogan, the podcaster in an 299 00:13:52,650 --> 00:13:56,580 S1: AI generated clip, is debating Ratatouille with Ben Shapiro, like 300 00:13:56,580 --> 00:13:57,720 S1: another conservative. 301 00:13:58,320 --> 00:14:01,110 S9: I just think I would love to get Ratatouille. Ratatouille, 302 00:14:01,140 --> 00:14:02,970 S9: Like have a little guy up there excuse me, you know, 303 00:14:02,970 --> 00:14:04,710 S9: making me cook delicious meals. No, I. 304 00:14:04,710 --> 00:14:06,300 S10: Understand. I've seen the film. You wouldn't. 305 00:14:06,300 --> 00:14:08,940 S9: Want that. A little guy pulling your hair, making you cook. 306 00:14:08,940 --> 00:14:09,179 S9: It's a. 307 00:14:09,179 --> 00:14:10,170 S10: Ridiculous concept. 308 00:14:10,170 --> 00:14:11,730 S9: You're telling me you wouldn't want that? 309 00:14:11,730 --> 00:14:13,140 S10: I just don't know why you're bringing it up. As 310 00:14:13,140 --> 00:14:14,910 S10: if it's something that could actually happen. 311 00:14:14,910 --> 00:14:17,730 S9: You wouldn't want a little Ratatouille guy up there doing whatever. 312 00:14:17,730 --> 00:14:19,530 S10: It's honestly insane that a rat would even be able 313 00:14:19,530 --> 00:14:20,490 S10: to cook in the first place. 314 00:14:20,490 --> 00:14:22,080 S9: He doesn't have to be cooking in the movie. 315 00:14:22,080 --> 00:14:22,620 S10: He's cooking. 316 00:14:22,620 --> 00:14:25,590 S9: Yeah, but in this scenario, though, he could do other stuff. Yes. 317 00:14:25,590 --> 00:14:27,150 S1: It's so funny. And it's like. And I think the 318 00:14:27,150 --> 00:14:30,450 S1: fact that it was so believable was incredible. 319 00:14:30,690 --> 00:14:33,780 S2: That was so good because like. And like, that's another thing, right? 320 00:14:33,780 --> 00:14:37,410 S2: Like the characterization of Ben Shapiro in these AI videos 321 00:14:37,410 --> 00:14:39,540 S2: is so good. Like, I'm sorry, I know this is 322 00:14:39,540 --> 00:14:41,400 S2: not talking about the issue, but it's just it's hilarious 323 00:14:41,400 --> 00:14:43,140 S2: to see how people choose to represent totally. 324 00:14:43,140 --> 00:14:45,540 S1: And I think that still falls into the category of 325 00:14:45,540 --> 00:14:49,440 S1: like more humorous than nefarious. But then very quickly, like 326 00:14:49,440 --> 00:14:53,550 S1: a quick comp to that is Emma Watson from Harry 327 00:14:53,550 --> 00:14:56,220 S1: Potter and various other things. Her voice was simulated to 328 00:14:56,220 --> 00:14:59,490 S1: have a reading out of Hitler's Mein Kampf, which you 329 00:14:59,490 --> 00:15:02,850 S1: can sort of see if you didn't understand the context 330 00:15:02,850 --> 00:15:05,550 S1: or if that was originally popped up on Reddit. And 331 00:15:05,550 --> 00:15:07,350 S1: then you saw that on Twitter, you would think, what 332 00:15:07,350 --> 00:15:10,530 S1: the hell is going on? Yeah. I mean, do you 333 00:15:10,530 --> 00:15:14,340 S1: think we're heading into a political election campaign in the 334 00:15:14,340 --> 00:15:18,570 S1: States where there's already so much concern about fake news 335 00:15:18,570 --> 00:15:22,979 S1: and misinformation? Like what does something like this add to that? 336 00:15:23,460 --> 00:15:25,380 S2: I think it adds like a layer of danger, right? Like, 337 00:15:25,530 --> 00:15:28,800 S2: because like, you know, as we well know, the initial 338 00:15:28,800 --> 00:15:33,150 S2: untruth often spreads further than the actual correction. Right? You know, 339 00:15:33,150 --> 00:15:35,730 S2: imagine if these deepfakes had existed when Donald Trump was 340 00:15:35,730 --> 00:15:38,790 S2: in office. You know, he was constantly saying things that he, 341 00:15:38,790 --> 00:15:40,800 S2: like very obviously said were not true. 342 00:15:40,800 --> 00:15:42,450 S1: And yeah, how could you tell the difference between. 343 00:15:42,450 --> 00:15:44,400 S2: Yeah, it gives it gives people like him more cover 344 00:15:44,400 --> 00:15:47,340 S2: to sort of, you know, perpetuate untruths or push propaganda. 345 00:15:47,370 --> 00:15:48,900 S2: You know, he could he could be reading from Mein 346 00:15:48,900 --> 00:15:50,730 S2: Kampf and then he could be like, oh, actually, that 347 00:15:50,730 --> 00:15:53,010 S2: was fake. Yeah, that's really interesting. And, you know, I 348 00:15:53,010 --> 00:15:55,979 S2: feel like there is a real danger of that becoming 349 00:15:55,980 --> 00:15:59,340 S2: a problem. And like that is that's concerning, Rob, because like, 350 00:15:59,340 --> 00:16:02,080 S2: you know, there's like there are real dangers in, you know, 351 00:16:02,130 --> 00:16:05,940 S2: their policies and their positions. And it creates like this 352 00:16:05,940 --> 00:16:08,370 S2: sort of environment where everyone else will be playing catch up, 353 00:16:08,370 --> 00:16:11,250 S2: you know, trying to dispel their own truths or sort 354 00:16:11,250 --> 00:16:13,620 S2: the fact from fiction. So I think there are real 355 00:16:13,620 --> 00:16:15,960 S2: like real issues with it. 356 00:16:15,960 --> 00:16:19,320 S1: I think the most interesting thing here is what you 357 00:16:19,320 --> 00:16:23,130 S1: are willing to consider is possible or viable out of this. 358 00:16:23,130 --> 00:16:26,040 S1: Like if if Joe, if a video gets released of 359 00:16:26,040 --> 00:16:30,630 S1: Joe Biden reading account, for example, it's like that is 360 00:16:30,630 --> 00:16:33,870 S1: so obviously like a deep fake that it's not really 361 00:16:33,870 --> 00:16:36,240 S1: much of a threat to his political career. But as 362 00:16:36,240 --> 00:16:38,160 S1: you mentioned, when someone like Donald Trump, it's like you 363 00:16:38,160 --> 00:16:40,080 S1: don't know if that could be real or not. So 364 00:16:40,080 --> 00:16:43,230 S1: maybe the debate over like deepfakes and what sort of 365 00:16:43,230 --> 00:16:46,170 S1: impact that could have on our perception of politics or 366 00:16:46,170 --> 00:16:48,240 S1: on campaigns is actually a bit of a fig leaf 367 00:16:48,240 --> 00:16:52,560 S1: for just like pointing to the kind of deeper risks 368 00:16:52,560 --> 00:16:54,900 S1: and challenges of having certain candidates like you should. You 369 00:16:54,900 --> 00:16:57,480 S1: should imagine in a world even with deepfakes, that if 370 00:16:57,480 --> 00:17:00,630 S1: someone says something so out there, that's obviously fake. But 371 00:17:00,630 --> 00:17:02,430 S1: then I wonder, like, what about those instances in the 372 00:17:02,430 --> 00:17:05,609 S1: middle where it is not Joe Biden reading from Mein Kampf, 373 00:17:05,609 --> 00:17:10,950 S1: but it is him announcing the placement of military warships 374 00:17:11,400 --> 00:17:14,550 S1: in an area that is contentious. So it's possible, but 375 00:17:14,850 --> 00:17:17,070 S1: he hasn't actually done it. But a video circulates that. 376 00:17:17,070 --> 00:17:19,590 S1: He's done it and then another country reacts. 377 00:17:19,770 --> 00:17:21,480 S2: I don't know if we're going to get like a 378 00:17:21,480 --> 00:17:24,119 S2: Cuban missile crisis out of deepfakes. Like is that sort 379 00:17:24,119 --> 00:17:24,990 S2: of what I'm. 380 00:17:24,990 --> 00:17:27,060 S1: Just I'm asking the question. I'm just asking questions here 381 00:17:28,380 --> 00:17:29,639 S1: not to go to Tucker Carlson. 382 00:17:29,640 --> 00:17:31,200 S2: Yeah, yeah. So true. But like, why not? 383 00:17:31,200 --> 00:17:32,810 S1: Why wouldn't we get that? Like if those videos. That's true. 384 00:17:32,850 --> 00:17:35,070 S1: These guys are talking about. Yeah, if you don't look 385 00:17:35,070 --> 00:17:37,560 S1: at that video and you just listen to Donald Trump 386 00:17:37,560 --> 00:17:40,560 S1: and Joe Biden arguing about Kendrick Lamar, it sounds like 387 00:17:40,560 --> 00:17:42,870 S1: Donald Trump and Joe Biden arguing about Kendrick Lamar. 388 00:17:42,869 --> 00:17:45,360 S2: Yeah, Yeah. And I think that's the thing, right? It's 389 00:17:45,359 --> 00:17:48,210 S2: like the the way the voices sound has gotten so 390 00:17:48,210 --> 00:17:50,850 S2: like scarily good in such a short space of time. 391 00:17:50,850 --> 00:17:53,100 S2: Like there's a noticeable difference between the early ones and 392 00:17:53,100 --> 00:17:55,619 S2: the and the late ones. I think the first one 393 00:17:55,619 --> 00:17:59,940 S2: I saw was Joe Biden saying like a whole bunch 394 00:17:59,940 --> 00:18:03,960 S2: of unprintable things about, you know, smoking weed. Right. And 395 00:18:03,960 --> 00:18:06,840 S2: he's talking about this specific brand of like Nubian Kush 396 00:18:06,840 --> 00:18:09,480 S2: or whatever. It's just some absolutely outlandish thing. And he 397 00:18:09,750 --> 00:18:11,969 S2: says a bunch of, like, awful and racist things. But 398 00:18:11,970 --> 00:18:14,790 S2: he also like like the idea is to mimic like 399 00:18:14,790 --> 00:18:17,159 S2: a specific type of person. Right? And like, sort of like, 400 00:18:17,160 --> 00:18:19,500 S2: put it out and like, it was so stilted and 401 00:18:19,500 --> 00:18:22,739 S2: unclear that like, obviously it was a deepfake, but it 402 00:18:22,740 --> 00:18:24,210 S2: was still weird to hear his voice saying it. But 403 00:18:24,210 --> 00:18:25,890 S2: now we're getting to the point where they actually sound 404 00:18:25,890 --> 00:18:29,969 S2: like they're having real conversations. You know, the intonations, the 405 00:18:29,970 --> 00:18:32,160 S2: way they interrupt each other, You know, what's to say 406 00:18:32,160 --> 00:18:33,689 S2: that this conversation isn't a deepfake? 407 00:18:34,020 --> 00:18:35,879 S1: Well, that's the great reveal of this, because it turns 408 00:18:35,880 --> 00:18:38,940 S1: out neither a surreal it's generated. Yeah. And I guess 409 00:18:38,940 --> 00:18:40,500 S1: I guess the other thing and you were alluding to 410 00:18:40,500 --> 00:18:42,770 S1: this before in terms of trust in media like. Where 411 00:18:42,770 --> 00:18:45,050 S1: these things emerge. Like if Joe Biden is going to 412 00:18:45,050 --> 00:18:48,560 S1: put missile cruise ship somewhere, he's not going to announce 413 00:18:48,560 --> 00:18:50,480 S1: that on TikTok or on like Discord. 414 00:18:50,510 --> 00:18:53,780 S2: No, no. But I mean, the so the Pokemon one, 415 00:18:53,780 --> 00:18:56,330 S2: which is unprintable. Sorry to keep bringing it up because 416 00:18:56,330 --> 00:18:57,350 S2: I'm sure everyone's really. 417 00:18:58,010 --> 00:18:59,750 S1: You can Google it if you want, but we're not 418 00:18:59,750 --> 00:18:59,960 S1: going to. 419 00:18:59,960 --> 00:19:01,280 S2: Talk about it. I'm not even going to give you 420 00:19:01,280 --> 00:19:03,710 S2: the information to Google it. That came with a video 421 00:19:03,710 --> 00:19:05,380 S2: that looked like him speaking at a rally. You know, 422 00:19:05,390 --> 00:19:08,780 S2: and when these when those videos like the real ones, 423 00:19:08,780 --> 00:19:11,180 S2: actually make it to TikTok, like that's how they look, right? 424 00:19:11,180 --> 00:19:12,709 S2: They look a little bit artifact. They look a little 425 00:19:12,710 --> 00:19:16,939 S2: bit weird and they still make it onto the platform. 426 00:19:16,940 --> 00:19:18,950 S2: So it wouldn't necessarily be like he's announcing it, but 427 00:19:18,950 --> 00:19:20,610 S2: it will look like it had been announced somewhere else. Right. 428 00:19:20,660 --> 00:19:23,780 S2: And the problem is, do people actually verify things? I mean, like, 429 00:19:23,780 --> 00:19:27,170 S2: say the TikTok says Joe Biden is putting, you know, 430 00:19:27,260 --> 00:19:31,160 S2: warships in, say, the South China Sea. I'm sure Jinping 431 00:19:31,160 --> 00:19:33,020 S2: is not going to be like, I saw us on TikTok. 432 00:19:33,020 --> 00:19:34,880 S2: I'm going to go see like he's like he's going 433 00:19:34,880 --> 00:19:37,070 S2: to like, you will call the the ambassador or whatever. 434 00:19:37,070 --> 00:19:40,100 S2: Like it's I don't think in those situations it's as 435 00:19:40,100 --> 00:19:41,690 S2: much of a clear problem as it is like sort 436 00:19:41,690 --> 00:19:44,630 S2: of this like lifting the baseline of sort of like 437 00:19:44,630 --> 00:19:47,450 S2: tension within, like a, like a country. Right. Is probably 438 00:19:47,750 --> 00:19:50,960 S2: more the propaganda usage of it than it is something else. Yeah. 439 00:19:50,960 --> 00:19:54,830 S1: No, fair enough. I feel like I'm less optimistic than 440 00:19:54,830 --> 00:19:58,250 S1: you are on that front. Yeah. I just think I 441 00:19:58,250 --> 00:20:00,500 S1: just think like in Pakistan right now, there's a huge 442 00:20:00,500 --> 00:20:05,119 S1: political crisis because these audiotapes have been leaked, allegedly with 443 00:20:05,359 --> 00:20:08,720 S1: former Prime Minister Imran Khan on them talking about corruption. 444 00:20:08,720 --> 00:20:11,810 S1: Apparently some stuff to do with him having affairs. And 445 00:20:11,810 --> 00:20:13,400 S1: it's like you removed the video of it and it's 446 00:20:13,400 --> 00:20:16,820 S1: just audio that seems great in an election campaign context. 447 00:20:16,820 --> 00:20:19,250 S1: Like you said it, you know, it just pops up 448 00:20:19,250 --> 00:20:21,290 S1: somewhere and a million people might share it. He might 449 00:20:21,290 --> 00:20:24,140 S1: issue a denial and I might even later out be 450 00:20:24,140 --> 00:20:27,530 S1: publicly recognized by some independent agency. This was a deep fake. 451 00:20:27,530 --> 00:20:30,470 S1: That's still like millions of people who heard that first thing. 452 00:20:30,470 --> 00:20:33,050 S2: Yeah, that's a good point. And I mean, in a 453 00:20:33,050 --> 00:20:37,219 S2: domestic election campaign, that's definitely a few notches less of 454 00:20:37,220 --> 00:20:40,790 S2: importance than, say, like a two nuclear, nuclear armed states 455 00:20:40,790 --> 00:20:42,470 S2: going at each other. But, you know, that could be 456 00:20:42,470 --> 00:20:44,670 S2: a problem in some way, like India and Pakistan, where 457 00:20:44,720 --> 00:20:48,020 S2: like they are two new nuclear armed states that are very, 458 00:20:48,020 --> 00:20:50,540 S2: you know, there's quite a lot of tension there and 459 00:20:50,900 --> 00:20:54,859 S2: something happening there could be very, very that's very scary. 460 00:20:54,859 --> 00:20:57,110 S2: That's a very scary and very important thing to sort 461 00:20:57,109 --> 00:21:00,020 S2: of point out, that the world is larger than our 462 00:21:00,020 --> 00:21:02,570 S2: little backyard and like there could be implications down the 463 00:21:02,570 --> 00:21:05,420 S2: line there for sure. And I think it absolutely would 464 00:21:05,420 --> 00:21:07,850 S2: have the power to swing domestic election campaign. I'm not 465 00:21:07,850 --> 00:21:09,919 S2: sure or maybe not swing, but like affect it in 466 00:21:09,920 --> 00:21:10,609 S2: some way. It affects it. 467 00:21:10,650 --> 00:21:13,520 S1: At least. Yeah. I think it's interesting that it took 468 00:21:13,520 --> 00:21:16,310 S1: me watching these hilarious TikTok videos to think about it. 469 00:21:16,310 --> 00:21:20,330 S1: And I just wonder, like it'll take something significant happening 470 00:21:20,330 --> 00:21:23,060 S1: before everyone realizes, Oh wow, we should really have kind 471 00:21:23,060 --> 00:21:25,220 S1: of figured out how to deal with this. Yeah. The 472 00:21:25,220 --> 00:21:27,200 S1: other thing I want to get your perspective on before 473 00:21:27,200 --> 00:21:31,490 S1: we wrap up this really interesting, slightly terrifying conversation is 474 00:21:31,490 --> 00:21:34,310 S1: in terms of the like, there's a debate within the 475 00:21:34,310 --> 00:21:38,210 S1: arts community about the role of artificial intelligence. Like a 476 00:21:38,210 --> 00:21:41,449 S1: lot of artists I know, and our colleague Liz Flux, 477 00:21:41,450 --> 00:21:44,149 S1: who's arts editor here at The Age she's written a 478 00:21:44,150 --> 00:21:47,810 S1: fair bit about this. You know, AI is another tool, right? 479 00:21:47,810 --> 00:21:51,080 S1: It can be used to create new work. It can 480 00:21:51,080 --> 00:21:55,100 S1: be used to help you use technology in a certain way, 481 00:21:55,100 --> 00:21:56,990 S1: in a way that all sorts of technologies used to 482 00:21:56,990 --> 00:21:59,959 S1: create art these days. But then there's another side to 483 00:21:59,960 --> 00:22:03,770 S1: this argument that says, well, no, these. Kind of tools 484 00:22:03,770 --> 00:22:08,239 S1: are derivative. They steal art, they devalue art by their 485 00:22:08,240 --> 00:22:13,490 S1: ability to create so much stuff of unknown quality so quickly. 486 00:22:13,970 --> 00:22:15,709 S1: Where do you sit on this or what do you 487 00:22:15,710 --> 00:22:16,940 S1: think about this debate? 488 00:22:17,540 --> 00:22:21,350 S2: Yeah, look, I don't think artists will ever like a 489 00:22:21,350 --> 00:22:24,580 S2: human artist will ever fully be irrelevant. I think like 490 00:22:24,590 --> 00:22:29,510 S2: that should be fairly uncontroversial. There is something like beautiful 491 00:22:29,510 --> 00:22:31,970 S2: about like expression that springs from nothing wrong that is 492 00:22:31,970 --> 00:22:34,219 S2: made with your own hands or whichever tools you have 493 00:22:34,220 --> 00:22:37,700 S2: available and sort of like, you know, removing the steps 494 00:22:37,700 --> 00:22:40,670 S2: in between your fingers and like, you know, like actually 495 00:22:40,670 --> 00:22:44,119 S2: making a beat on in Pro Tools or whatever is 496 00:22:44,119 --> 00:22:46,580 S2: like removing some of that removes some of the magic, right? 497 00:22:46,730 --> 00:22:48,560 S2: And some of the decisions and the choices you make, 498 00:22:48,560 --> 00:22:51,290 S2: which is like a part of like creation of art, right? 499 00:22:51,740 --> 00:22:55,700 S2: But I don't necessarily think that it will be great 500 00:22:55,700 --> 00:22:59,450 S2: for artists either, because artists, as we all know, like generally, 501 00:22:59,450 --> 00:23:02,480 S2: unless they're like, you know, touring the globe, doing a 502 00:23:02,570 --> 00:23:06,410 S2: sold out shows or selling art into that, like visual art. 503 00:23:06,410 --> 00:23:09,050 S2: And so like those weird collectors sort of circles, like 504 00:23:09,050 --> 00:23:11,980 S2: there's unless you're one of those people, you don't typically yeah, 505 00:23:12,050 --> 00:23:14,180 S2: you're not really well compensated. So there is like a 506 00:23:14,180 --> 00:23:18,860 S2: bit of concern that I share within my my circles 507 00:23:18,859 --> 00:23:22,640 S2: like that, like artists will just sort of no longer 508 00:23:22,640 --> 00:23:25,040 S2: be able to actually get paid for the sort of 509 00:23:25,040 --> 00:23:26,899 S2: mega things I do get paid for. So say you're 510 00:23:26,900 --> 00:23:30,850 S2: an artist writing jingles for like for ads, you know, 511 00:23:30,859 --> 00:23:33,649 S2: that is a way you can make money on the 512 00:23:33,650 --> 00:23:35,480 S2: side when you might be doing something that you'd rather 513 00:23:35,480 --> 00:23:38,689 S2: be doing right. You know, expressing yourself more freely. If 514 00:23:38,690 --> 00:23:41,570 S2: a big company has the ability to generate that in 515 00:23:41,570 --> 00:23:46,280 S2: the like or generate hundreds of options within a few moments, 516 00:23:46,280 --> 00:23:48,379 S2: they don't really have the need to pay like a 517 00:23:48,380 --> 00:23:52,850 S2: suffering artist to create something like something that is, unless 518 00:23:52,850 --> 00:23:55,810 S2: they're looking for a bespoke solution, right? So, you know, 519 00:23:55,880 --> 00:23:58,790 S2: that's of concern, like sort of closes the avenues for 520 00:23:58,790 --> 00:24:01,159 S2: artists to make a living, otherwise offer their art and 521 00:24:01,340 --> 00:24:05,869 S2: keep practicing and also there's a question of concentrating the 522 00:24:05,869 --> 00:24:08,240 S2: ability to make art within the hands of a few. 523 00:24:08,240 --> 00:24:10,340 S2: So say you're one of these companies who have made 524 00:24:10,940 --> 00:24:14,480 S2: these AI art generators. You then wield an enormous amount 525 00:24:14,480 --> 00:24:16,760 S2: of power because you can create as much art as 526 00:24:16,760 --> 00:24:19,370 S2: you like. You hold the means of production, as it were, 527 00:24:19,369 --> 00:24:24,109 S2: to to actually create these things. And I think history 528 00:24:24,109 --> 00:24:27,830 S2: will show that the the consolidation and the monopolization of 529 00:24:27,830 --> 00:24:33,650 S2: of any sort of platform or or tool is generally bad. 530 00:24:33,920 --> 00:24:37,040 S2: It doesn't necessarily create an environment where there is lots 531 00:24:37,040 --> 00:24:39,710 S2: of innovation or there is lots of boundary pushing, genre 532 00:24:39,710 --> 00:24:40,639 S2: pushing ideas. 533 00:24:40,760 --> 00:24:42,679 S1: That's smart. I think that's right to think about it 534 00:24:42,680 --> 00:24:46,640 S1: from the perspective of who controls these things and to 535 00:24:46,640 --> 00:24:49,070 S1: what end are they using them for? Like, there is 536 00:24:49,070 --> 00:24:52,250 S1: a world in which this kind of technology is used 537 00:24:52,250 --> 00:24:55,580 S1: to liberate people from very boring jobs and tasks and 538 00:24:55,580 --> 00:24:59,600 S1: to help them create exciting new stuff. But under our 539 00:24:59,600 --> 00:25:04,430 S1: current sort of structures and economic situations, they'll overwhelmingly be 540 00:25:04,430 --> 00:25:08,510 S1: used to minimize costs by laying people off and extract 541 00:25:08,510 --> 00:25:10,940 S1: as much value and revenue for the people who own 542 00:25:10,940 --> 00:25:12,139 S1: them as possible. 543 00:25:12,140 --> 00:25:14,570 S2: Yeah, and I think, you know, you don't even have 544 00:25:14,570 --> 00:25:16,369 S2: to go far back in history to actually see an 545 00:25:16,369 --> 00:25:19,280 S2: example of that happening. Right. Like the dawn of the computer. Exactly. 546 00:25:19,280 --> 00:25:21,950 S2: Supposed to be like, oh, we'll all work for four 547 00:25:21,950 --> 00:25:23,810 S2: day weeks, you know, three hours a day, and that'll 548 00:25:23,840 --> 00:25:25,100 S2: be it. You'll be able to log off and go home. 549 00:25:25,100 --> 00:25:26,000 S2: But that's not what happened. 550 00:25:26,000 --> 00:25:27,650 S1: No, we're working 12 hours a day. Seven days. 551 00:25:27,650 --> 00:25:30,050 S2: A week. Yeah. And it's getting worse, right? And and 552 00:25:30,050 --> 00:25:33,470 S2: I think that's the that makes me very pessimistic about 553 00:25:33,470 --> 00:25:38,270 S2: the application within businesses of AI art and AI more generally. Right. 554 00:25:38,630 --> 00:25:42,740 S2: But then there's also on the positive side, like it 555 00:25:42,740 --> 00:25:46,580 S2: lowers the bar for people to enter artistry, right? Like 556 00:25:46,580 --> 00:25:48,619 S2: there's no way in hell I could make a beat 557 00:25:48,619 --> 00:25:52,010 S2: that quickly and come up with like a four bar rhyme. 558 00:25:52,100 --> 00:25:55,160 S2: And I think that, you know, lowering the bar for that, 559 00:25:55,160 --> 00:25:57,859 S2: like it gives people who have ideas but maybe not 560 00:25:57,859 --> 00:26:00,619 S2: necessarily the skills, the ability to express themselves. And I 561 00:26:00,619 --> 00:26:04,520 S2: think that will create and that's not even going into like, 562 00:26:04,520 --> 00:26:07,730 S2: you know, the AI generated rappers which are mired in controversy, right? 563 00:26:07,730 --> 00:26:12,619 S2: But like, I'm just saying, like some disembodied and, you know, 564 00:26:12,619 --> 00:26:15,409 S2: sort of like anonymous voice, you know, gives you the 565 00:26:15,410 --> 00:26:18,020 S2: opportunity to sort of create something that may not necessarily 566 00:26:18,020 --> 00:26:21,080 S2: be like a Top 40 hit or something that's going 567 00:26:21,080 --> 00:26:23,960 S2: to change the way we like the way we think 568 00:26:23,960 --> 00:26:27,619 S2: about the world. Like, say, like Pimp a Butterfly did once, 569 00:26:27,619 --> 00:26:30,560 S2: you know, like those things may not be possible yet 570 00:26:30,560 --> 00:26:34,310 S2: or at all with AI generation, but it does give 571 00:26:34,310 --> 00:26:37,070 S2: people who lack the skills like the the ability to 572 00:26:37,070 --> 00:26:39,409 S2: get started or to at least express themselves in a 573 00:26:39,410 --> 00:26:41,300 S2: way that they wish to. So I don't necessarily think 574 00:26:41,300 --> 00:26:43,250 S2: that that's a terrible thing. I think that's I think 575 00:26:43,250 --> 00:26:46,430 S2: it's quite interesting. It's worth considering. But unfortunately, we don't 576 00:26:46,430 --> 00:26:49,700 S2: get to pick and choose which social changes we get. 577 00:26:49,790 --> 00:26:54,230 S1: Totally. That's a really interesting and succinct place to end 578 00:26:54,230 --> 00:26:56,899 S1: a conversation that is probably really just starting when you 579 00:26:56,900 --> 00:26:58,940 S1: think about how we as a society across the globe 580 00:26:58,940 --> 00:27:01,220 S1: are thinking about this stuff. Jim Malloy, thank you so 581 00:27:01,220 --> 00:27:02,359 S1: much for your time today. 582 00:27:02,380 --> 00:27:02,950 UU: Thanks for having me. 583 00:27:07,440 --> 00:27:10,050 S1: This episode of The Drop was produced by Kai Wang. 584 00:27:10,380 --> 00:27:13,260 S1: If you enjoyed listening to today's episode of The Drop, 585 00:27:13,260 --> 00:27:15,970 S1: make sure to follow us in your favorite podcast app. 586 00:27:15,990 --> 00:27:18,210 S1: Give us a review or better yet, share it with 587 00:27:18,210 --> 00:27:20,340 S1: a friend. I'm Osman Farooqui. 588 00:27:20,369 --> 00:27:21,030 UU: See you soon.