WEBVTT - how an editor uncovered an AI journalist

0:00:10.960 --> 0:00:12.960
<v Speaker 1>I hadn't done a pitch call out about a year,

0:00:12.960 --> 0:00:15.120
<v Speaker 1>I guess, so I hadn't really been paying attention to

0:00:15.200 --> 0:00:17.759
<v Speaker 1>what that world was like. But this time I got like,

0:00:17.960 --> 0:00:22.000
<v Speaker 1>way more pitches than ever, very quickly, from all over

0:00:22.040 --> 0:00:24.680
<v Speaker 1>the place. It was very different than the year before.

0:00:25.200 --> 0:00:28.280
<v Speaker 2>Nicholas hun Brown is the editor of The Local, which

0:00:28.360 --> 0:00:32.120
<v Speaker 2>is an independent online magazine that covers social issues in Toronto.

0:00:32.600 --> 0:00:35.360
<v Speaker 2>Last September, they put out an open call for freelance

0:00:35.440 --> 0:00:38.320
<v Speaker 2>journalists to pitch stories for their upcoming issue.

0:00:38.400 --> 0:00:41.760
<v Speaker 1>We were looking to assign stories about healthcare and money.

0:00:41.880 --> 0:00:45.839
<v Speaker 1>There's been some creeping privatization in Ontario where we live,

0:00:46.159 --> 0:00:48.600
<v Speaker 1>so we put out a call for pitches. I posted

0:00:48.600 --> 0:00:51.880
<v Speaker 1>something on Blue Sky just asking freelancers give me a suggestions.

0:00:52.200 --> 0:00:54.920
<v Speaker 2>Nick got a lot of pitches, but among those, there

0:00:55.000 --> 0:00:57.560
<v Speaker 2>was one in particular that stood out.

0:00:57.800 --> 0:01:00.640
<v Speaker 1>This was a writer who called herself Victoria Goald. She

0:01:00.720 --> 0:01:02.520
<v Speaker 1>said in the email that she'd written for a bunch

0:01:02.560 --> 0:01:05.160
<v Speaker 1>of these Canadian publications that do similar work to what

0:01:05.200 --> 0:01:07.319
<v Speaker 1>we do. It was a well written pitch, I thought,

0:01:07.480 --> 0:01:09.160
<v Speaker 1>and when I did a quick Google of her, she

0:01:09.240 --> 0:01:12.320
<v Speaker 1>had bylines and a bunch of you know, repude of publications,

0:01:12.360 --> 0:01:15.319
<v Speaker 1>The Guardian, New York Magazine, places like that. You know,

0:01:15.360 --> 0:01:17.680
<v Speaker 1>when you see someone has written for all these other publications,

0:01:17.880 --> 0:01:20.399
<v Speaker 1>my first instinct was, like, that seems legit. You know,

0:01:20.440 --> 0:01:22.840
<v Speaker 1>these other publications are good publications. I'm not going to

0:01:22.840 --> 0:01:23.640
<v Speaker 1>be too suspicious.

0:01:24.080 --> 0:01:24.720
<v Speaker 3>Why would you be?

0:01:25.080 --> 0:01:28.840
<v Speaker 2>Yeah, but Victoria Goldie turned out not to be who

0:01:28.880 --> 0:01:31.760
<v Speaker 2>she said she was, and this sent Nick down a

0:01:31.800 --> 0:01:35.839
<v Speaker 2>really weird rabbit hole investigation. You've heard of fake news, Well,

0:01:36.040 --> 0:01:44.200
<v Speaker 2>welcome to a new scam. Fake journalists from Kaleidoscope and

0:01:44.280 --> 0:01:45.240
<v Speaker 2>iHeart Podcasts.

0:01:50.560 --> 0:01:56.560
<v Speaker 3>This is kill Switch. I'm Dexterah Thomas.

0:01:58.040 --> 0:02:35.359
<v Speaker 4>I'm sorry, I'm goodbye.

0:02:39.240 --> 0:02:41.440
<v Speaker 2>Back in September, when Nick first put out this call

0:02:41.480 --> 0:02:45.000
<v Speaker 2>for pitches, the magazine was deliberately trying to expand their

0:02:45.040 --> 0:02:48.399
<v Speaker 2>freelance writer pool. They were hoping to find new writers,

0:02:48.680 --> 0:02:51.600
<v Speaker 2>new voices, people with interesting stories to tell.

0:02:52.280 --> 0:02:53.880
<v Speaker 1>Part of our mandate is we want to work with

0:02:53.919 --> 0:02:56.560
<v Speaker 1>young writers. We want to work with people from different

0:02:56.560 --> 0:02:59.720
<v Speaker 1>parts of the city, different backgrounds. We are actively seeking

0:02:59.760 --> 0:03:02.760
<v Speaker 1>to work with people who maybe don't have the clips,

0:03:02.880 --> 0:03:05.960
<v Speaker 1>are not as experienced, but are able to bring a

0:03:06.000 --> 0:03:07.880
<v Speaker 1>point of view and a story to us that we

0:03:07.919 --> 0:03:10.800
<v Speaker 1>won't found elsewhere, So like the pitch meant a whole lot,

0:03:10.840 --> 0:03:12.920
<v Speaker 1>Like if you could tell me an awesome story in

0:03:12.960 --> 0:03:15.080
<v Speaker 1>a pitch, if I could see your style on the pitch,

0:03:15.160 --> 0:03:17.440
<v Speaker 1>if I could see that you had like an idea

0:03:17.480 --> 0:03:19.880
<v Speaker 1>I couldn't find elsewhere, that would go a long way.

0:03:21.120 --> 0:03:24.320
<v Speaker 2>And Victoria Goldie's pitch hit all those things. It was

0:03:24.440 --> 0:03:27.359
<v Speaker 2>exactly what Nick was looking for. Her pitch was about

0:03:27.360 --> 0:03:31.400
<v Speaker 2>a debate in Canada about something called membership medicine. So

0:03:31.560 --> 0:03:34.080
<v Speaker 2>you probably know that Canada has free essential health care,

0:03:34.120 --> 0:03:38.040
<v Speaker 2>but membership medicine is basically when patients pay their doctors

0:03:38.080 --> 0:03:42.840
<v Speaker 2>directly for more premium services, and man, Victoria was on it.

0:03:42.960 --> 0:03:46.160
<v Speaker 2>In that initial email to Nick, she wrote, quote, the

0:03:46.240 --> 0:03:50.040
<v Speaker 2>story would track how these plans transform healthcare into something

0:03:50.080 --> 0:03:53.720
<v Speaker 2>resembling Netflix or Amazon Prime, and what this means for

0:03:53.800 --> 0:03:57.760
<v Speaker 2>a public system that has long prided itself on universality.

0:03:58.160 --> 0:04:00.120
<v Speaker 1>The pitch she sent it indicated that she done a

0:04:00.120 --> 0:04:03.080
<v Speaker 1>lot of reporting already. She'd done some research. She said,

0:04:03.080 --> 0:04:06.400
<v Speaker 1>She'd spoken with a couple prominent doctors here. She seemed

0:04:06.400 --> 0:04:08.400
<v Speaker 1>to have some like patients who are willing to talk

0:04:08.440 --> 0:04:11.880
<v Speaker 1>about their experiences with privatized medical care. It seemed like

0:04:11.880 --> 0:04:15.040
<v Speaker 1>a promising pitch, so we thought, you know, let's take

0:04:15.040 --> 0:04:15.800
<v Speaker 1>a shot on this person.

0:04:17.000 --> 0:04:20.240
<v Speaker 2>Just as a fellow freelancer. I appreciate Nick taking that

0:04:20.240 --> 0:04:23.440
<v Speaker 2>shot on Victoria because I've been there. You're trying to

0:04:23.480 --> 0:04:25.920
<v Speaker 2>show an editor that you're the person to hire, so

0:04:26.040 --> 0:04:29.440
<v Speaker 2>you do all this extra work up front, like, look,

0:04:29.600 --> 0:04:32.120
<v Speaker 2>I've already done this interview, I've already got access to

0:04:32.160 --> 0:04:34.520
<v Speaker 2>the hospital. Just say the word boss, and I will

0:04:34.520 --> 0:04:39.000
<v Speaker 2>write the best article just for you. So please understand

0:04:39.080 --> 0:04:42.960
<v Speaker 2>right now, at this point, I am rooting for Victoria Goldie.

0:04:43.120 --> 0:04:46.000
<v Speaker 2>So in their next meeting, Nick and his editorial team

0:04:46.080 --> 0:04:48.880
<v Speaker 2>decide to move forward with Victoria. But then he looked

0:04:48.920 --> 0:04:53.320
<v Speaker 2>at Victoria's original email again and realized something that he

0:04:53.680 --> 0:04:54.920
<v Speaker 2>hadn't noticed before.

0:04:55.560 --> 0:04:57.560
<v Speaker 1>You know, I went back to her pitch, and some

0:04:57.680 --> 0:05:01.160
<v Speaker 1>other questions popped up in my mind, Like the biggest

0:05:01.160 --> 0:05:03.240
<v Speaker 1>one was is she in Toronto? You know, when I

0:05:03.240 --> 0:05:04.880
<v Speaker 1>looked at her Guardian stuff, it was all about her

0:05:05.000 --> 0:05:09.040
<v Speaker 1>being in England. Once you'd done stuff for American magazines,

0:05:09.080 --> 0:05:11.479
<v Speaker 1>it seemed as if she was writing as an American,

0:05:11.720 --> 0:05:13.440
<v Speaker 1>So I wondered, you know that there are people who

0:05:13.440 --> 0:05:16.360
<v Speaker 1>are international. There's no reason that you could be publishing

0:05:16.400 --> 0:05:19.000
<v Speaker 1>and reporting from different places. But yeah, it was a

0:05:19.040 --> 0:05:21.200
<v Speaker 1>question I had. And the second was that she'd done

0:05:21.279 --> 0:05:23.240
<v Speaker 1>so much pre reporting for this. She'd done so much

0:05:23.279 --> 0:05:26.200
<v Speaker 1>work on this pitch. Within like, you know, a week

0:05:26.279 --> 0:05:28.560
<v Speaker 1>or two of me setting out the call out, she'd

0:05:28.560 --> 0:05:31.200
<v Speaker 1>interviewed four people or something. We're in that first pitch.

0:05:31.400 --> 0:05:33.680
<v Speaker 1>And I was a freelancer forever. I know that process,

0:05:33.720 --> 0:05:36.640
<v Speaker 1>and sometimes it's useful to do some reporting before you

0:05:36.640 --> 0:05:39.400
<v Speaker 1>get a story assigned. But like that much is there's

0:05:39.440 --> 0:05:41.240
<v Speaker 1>a lot that's a risk, right, there's no guarantee that

0:05:41.240 --> 0:05:42.960
<v Speaker 1>you're going to land the story. So then I thought,

0:05:43.000 --> 0:05:45.119
<v Speaker 1>you know, there's times where someone could be a student

0:05:45.160 --> 0:05:48.440
<v Speaker 1>journalist and then they're reusing interviews they've done before. There's

0:05:48.440 --> 0:05:52.040
<v Speaker 1>different reasons that could happen. But that made me question

0:05:52.160 --> 0:05:56.280
<v Speaker 1>some things, so he dug a little deeper. So then

0:05:56.320 --> 0:05:59.200
<v Speaker 1>I wanted to just see what stories she'd done for

0:05:59.240 --> 0:06:01.840
<v Speaker 1>Canadian publics. I'd seen the work that she'd done at

0:06:01.839 --> 0:06:04.480
<v Speaker 1>The Guardian, the work she'd done elsewhere, and in her

0:06:04.480 --> 0:06:06.520
<v Speaker 1>pitch she said she'd written for our national in these

0:06:06.560 --> 0:06:08.880
<v Speaker 1>papers of the Globe and Mail, the Walrus kind of

0:06:09.080 --> 0:06:11.159
<v Speaker 1>magazines that are like us. So I just googled her

0:06:11.279 --> 0:06:14.880
<v Speaker 1>name along with those publications, and nothing came up. This

0:06:15.120 --> 0:06:18.640
<v Speaker 1>is where Nick starts to get really suspicious. Could Victoria

0:06:18.760 --> 0:06:21.680
<v Speaker 1>just be lying about publishing in all these different places?

0:06:22.160 --> 0:06:25.320
<v Speaker 1>And if she was, what else could she be lying about?

0:06:25.920 --> 0:06:27.680
<v Speaker 1>She went back to look at her pitch to see

0:06:27.680 --> 0:06:30.640
<v Speaker 1>if there was any other explanation for this. Then I

0:06:30.680 --> 0:06:32.400
<v Speaker 1>was looking at all these quotes that she had in

0:06:32.440 --> 0:06:35.239
<v Speaker 1>her pitch, and some of them were from regular patients,

0:06:35.240 --> 0:06:38.040
<v Speaker 1>she said, and some were from some prominent doctors in Toronto.

0:06:38.120 --> 0:06:40.279
<v Speaker 1>And my colleague happens to know one of these doctors.

0:06:40.320 --> 0:06:41.720
<v Speaker 1>So he shot her an email and said, did you

0:06:41.760 --> 0:06:44.880
<v Speaker 1>speak to a journalist called Victoria Goldie and this doctor?

0:06:44.960 --> 0:06:48.279
<v Speaker 1>Doctor Daniel Martin said, no, never heard of her. That's

0:06:48.880 --> 0:06:52.800
<v Speaker 1>very weird that I'm being quoted this way. So when

0:06:52.800 --> 0:06:55.560
<v Speaker 1>that happened, I kind of knew. I knew what was

0:06:55.600 --> 0:06:57.919
<v Speaker 1>going on, right, Like, I looked at the pitch again,

0:06:58.160 --> 0:07:00.760
<v Speaker 1>I read it with kind of different eyes, and it

0:07:00.760 --> 0:07:03.839
<v Speaker 1>felt to me as if someone had been using generative

0:07:03.880 --> 0:07:06.680
<v Speaker 1>AI to create a pitch with made up quotes from

0:07:06.760 --> 0:07:10.000
<v Speaker 1>real people, and they had been making up some of

0:07:10.040 --> 0:07:14.120
<v Speaker 1>their botolines in Canadian publications, and when you reread it again,

0:07:14.200 --> 0:07:18.160
<v Speaker 1>it was like there was some chat gptisms, you know,

0:07:18.480 --> 0:07:22.120
<v Speaker 1>if there was some like formulaic writing about this pitch

0:07:22.160 --> 0:07:24.320
<v Speaker 1>is important because of X, it's timely, because of why,

0:07:24.320 --> 0:07:26.480
<v Speaker 1>it's perfect for your publication, because of Z. It felt

0:07:26.480 --> 0:07:28.800
<v Speaker 1>a little When I looked at it more closely, you

0:07:28.840 --> 0:07:31.440
<v Speaker 1>could see some things that felt a little inhuman.

0:07:31.560 --> 0:07:35.320
<v Speaker 2>I guess, hmmm, See that's tough because I've definitely pitched

0:07:35.320 --> 0:07:39.520
<v Speaker 2>things that I've used language like that. No, totally, like

0:07:39.560 --> 0:07:42.280
<v Speaker 2>that's what you do, Like you tell, hey, I'd like

0:07:42.400 --> 0:07:45.080
<v Speaker 2>to write this, and it's timely because this is happening.

0:07:45.080 --> 0:07:47.400
<v Speaker 2>It's the anniversary of whatever. We should revisit this.

0:07:47.720 --> 0:07:50.840
<v Speaker 1>Yeah, no, all the telles for chat GPT or whatever.

0:07:50.960 --> 0:07:53.040
<v Speaker 1>Like people say looking for M dashes and I'm so

0:07:53.080 --> 0:07:55.000
<v Speaker 1>resentful of that. I use M dashes all the time,

0:07:55.120 --> 0:07:58.520
<v Speaker 1>like it's same. So I don't know, I don't have

0:07:58.560 --> 0:08:01.400
<v Speaker 1>a good way of spawning something. But the made up

0:08:01.480 --> 0:08:04.320
<v Speaker 1>quotes and the kind of made up bylines made me

0:08:04.320 --> 0:08:05.480
<v Speaker 1>think this is what was happening.

0:08:06.240 --> 0:08:09.400
<v Speaker 2>So okay, you're starting to think that this is AI.

0:08:09.960 --> 0:08:10.800
<v Speaker 2>What do you do next?

0:08:11.440 --> 0:08:13.560
<v Speaker 1>Then I just got really curious, right, Like, it's clear

0:08:13.600 --> 0:08:15.239
<v Speaker 1>that this was a fake pitch, but like, what about

0:08:15.240 --> 0:08:18.440
<v Speaker 1>all of those other bylines, Like all those other stories.

0:08:18.520 --> 0:08:21.080
<v Speaker 1>There's dozens of them across the Internet that she's written.

0:08:21.200 --> 0:08:23.360
<v Speaker 1>Like if she was faking stuff in this pitch for

0:08:23.440 --> 0:08:26.400
<v Speaker 1>us so blatantly, what about those stories?

0:08:28.760 --> 0:08:32.040
<v Speaker 2>So Nick started looking at Victoria Goldie's previous body of

0:08:32.120 --> 0:08:36.120
<v Speaker 2>work and he kind of got obsessed, and that obsession

0:08:36.720 --> 0:08:41.640
<v Speaker 2>led him to some pretty weird discoveries. That's after the break.

0:08:52.920 --> 0:08:55.680
<v Speaker 2>After Nick puts together that Victoria seems to have not

0:08:55.760 --> 0:08:59.240
<v Speaker 2>only chat GPT or pitch, but also lied about working

0:08:59.240 --> 0:09:02.480
<v Speaker 2>on other Canadians publications, he decides that he needs to

0:09:02.520 --> 0:09:05.880
<v Speaker 2>figure out how far this rabbit hole goes. He's pretty

0:09:05.960 --> 0:09:11.040
<v Speaker 2>sure he's being laed to, but he's still curious. You

0:09:11.080 --> 0:09:12.839
<v Speaker 2>didn't confront her immediately?

0:09:13.520 --> 0:09:15.800
<v Speaker 1>No, I wanted to know a little more, did. I

0:09:15.800 --> 0:09:17.959
<v Speaker 1>did send some follow up emails. I said, like, did

0:09:17.960 --> 0:09:19.679
<v Speaker 1>you actually speak with these people? And she said yes,

0:09:19.720 --> 0:09:21.400
<v Speaker 1>I did. I said, are you based in Toronto because

0:09:21.400 --> 0:09:23.640
<v Speaker 1>we need local reporting? She said, yeah, absolutely and based

0:09:23.640 --> 0:09:27.199
<v Speaker 1>in Toronto. So the kind of like I understood the

0:09:27.280 --> 0:09:29.000
<v Speaker 1>level of deception that was going on. I guess I

0:09:29.080 --> 0:09:31.640
<v Speaker 1>understood that she was making stuff up. So when I

0:09:31.720 --> 0:09:35.080
<v Speaker 1>started looking into the work that she'd published, the first

0:09:35.080 --> 0:09:37.520
<v Speaker 1>one I looked at was this feature from the Journal

0:09:37.720 --> 0:09:40.280
<v Speaker 1>of the Law Society of Scotland, which is like an

0:09:40.320 --> 0:09:43.320
<v Speaker 1>in depth feature that was about law firms disappearing from

0:09:43.960 --> 0:09:46.680
<v Speaker 1>the Scottish Highlands and like what were rural people doing?

0:09:46.720 --> 0:09:49.120
<v Speaker 1>And I quoted like members of parliament, I quoted a

0:09:49.120 --> 0:09:52.080
<v Speaker 1>bunch of lawyers, academics, and then a bunch of regular people.

0:09:52.200 --> 0:09:54.520
<v Speaker 1>So I was just wondering, like, this is a real article.

0:09:54.960 --> 0:09:56.760
<v Speaker 1>I couldn't find any of the regular people, you know,

0:09:56.760 --> 0:10:00.320
<v Speaker 1>when she begins an anecdote, like Fiona forty to a

0:10:00.320 --> 0:10:05.160
<v Speaker 1>school teacher in Glengarry or whatever, like those people google them,

0:10:05.200 --> 0:10:07.400
<v Speaker 1>can find any of them. So then I emailed the

0:10:07.440 --> 0:10:09.480
<v Speaker 1>people that she spoke to who were real, like a

0:10:09.520 --> 0:10:12.600
<v Speaker 1>professor called Elaine Sutherland who she quoted in that piece.

0:10:12.640 --> 0:10:14.040
<v Speaker 1>And I just emailed and said, did you speak to

0:10:14.040 --> 0:10:17.800
<v Speaker 1>this writer? And she said no, absolutely not. But what's

0:10:17.840 --> 0:10:19.960
<v Speaker 1>weird is that this sounds like the kind of thing

0:10:19.960 --> 0:10:22.280
<v Speaker 1>I would say, wow, And this is the response that

0:10:22.280 --> 0:10:24.959
<v Speaker 1>I was getting in a lot of places, and I

0:10:25.000 --> 0:10:26.880
<v Speaker 1>don't know, it's one of the eerier things. Right. It's

0:10:26.880 --> 0:10:29.280
<v Speaker 1>not someone just making stuff up, but it is, I

0:10:29.320 --> 0:10:33.559
<v Speaker 1>believe AI searching through the internet thinking who is the

0:10:33.679 --> 0:10:36.760
<v Speaker 1>likely person who can talk about Scottish law firms disappearing

0:10:36.760 --> 0:10:39.000
<v Speaker 1>from the Highlands, you find someone who maybe has spoken

0:10:39.040 --> 0:10:40.800
<v Speaker 1>too that in the past, or written a research paper

0:10:40.800 --> 0:10:43.160
<v Speaker 1>about it, or whatever it is, and then kind of

0:10:43.280 --> 0:10:46.960
<v Speaker 1>invents quotes that sound kind of like in the ballpark

0:10:47.000 --> 0:10:48.640
<v Speaker 1>of what you might say. So that was the first

0:10:48.679 --> 0:10:51.280
<v Speaker 1>piece that I thought, Okay, this is a made up story,

0:10:51.440 --> 0:10:53.320
<v Speaker 1>and then I just kept looking. There were more. There

0:10:53.360 --> 0:10:57.000
<v Speaker 1>was a story in Dwell that quoted like ten prominent

0:10:57.400 --> 0:11:00.800
<v Speaker 1>international designers and architects, and I just started sending emails

0:11:00.840 --> 0:11:03.360
<v Speaker 1>to all of them, and they had not spoken to

0:11:03.400 --> 0:11:03.840
<v Speaker 1>this person.

0:11:04.960 --> 0:11:08.679
<v Speaker 2>Victoria had published work in The Guardian, in Business Insider,

0:11:09.000 --> 0:11:13.600
<v Speaker 2>New York Magazine, Rolling Stone, Africa, Vogue, Philippines. I could

0:11:13.640 --> 0:11:16.800
<v Speaker 2>keep going here. This is a resume that almost any

0:11:16.840 --> 0:11:19.120
<v Speaker 2>freelance journalists would be really jealous of.

0:11:19.760 --> 0:11:21.800
<v Speaker 1>How many articles should you have out there? I think

0:11:21.840 --> 0:11:23.600
<v Speaker 1>over two dozen. I think I've got a spreadsheet of

0:11:23.679 --> 0:11:26.160
<v Speaker 1>over two dozen articles. You made a spreadsheet I got,

0:11:26.200 --> 0:11:29.080
<v Speaker 1>I got obsessed. I don't know I was not gonna

0:11:29.080 --> 0:11:31.120
<v Speaker 1>write about this. We're a local publication that only reads

0:11:31.120 --> 0:11:32.800
<v Speaker 1>about Toronto, right, I was just doing this set of

0:11:32.880 --> 0:11:35.720
<v Speaker 1>curiosity to begin with. But I don't know, it just

0:11:35.720 --> 0:11:38.720
<v Speaker 1>seems so wild to me that this stuff was out there.

0:11:38.840 --> 0:11:42.200
<v Speaker 1>And maybe it sound like overly, you know, people are

0:11:42.280 --> 0:11:45.120
<v Speaker 1>lying on the internet, but I found it shocking. I guess,

0:11:45.920 --> 0:11:48.400
<v Speaker 1>how dare somebody lie on the internet. Yeah, yeah, but

0:11:48.480 --> 0:11:50.800
<v Speaker 1>I lie on the Internet under the guardian's name or whatever.

0:11:50.960 --> 0:11:52.640
<v Speaker 3>That is different. Yeah, that is different.

0:11:52.720 --> 0:11:55.280
<v Speaker 2>Yeah, getting on there and saying, hey, yo, I don't know,

0:11:55.320 --> 0:11:58.800
<v Speaker 2>somebody like lying on their dating profile. Yo, I'm six

0:11:58.840 --> 0:12:00.880
<v Speaker 2>foot two and like, bro, you're five to three.

0:12:00.920 --> 0:12:01.480
<v Speaker 1>Like okay.

0:12:01.480 --> 0:12:04.480
<v Speaker 2>Maybe not a nice thing to do, but also very

0:12:04.480 --> 0:12:08.000
<v Speaker 2>different from fooling the public into thinking that something happened.

0:12:08.120 --> 0:12:10.320
<v Speaker 2>And some people said some things that they actually never did.

0:12:10.600 --> 0:12:11.240
<v Speaker 1>Yeah.

0:12:11.440 --> 0:12:14.480
<v Speaker 2>So one place that had published Victoria's work had actually

0:12:14.520 --> 0:12:17.280
<v Speaker 2>already taken some stuff down, but they didn't say it

0:12:17.320 --> 0:12:18.120
<v Speaker 2>was because of AI.

0:12:19.040 --> 0:12:22.280
<v Speaker 1>I emailed with editors of other publications and spoke to

0:12:22.320 --> 0:12:25.120
<v Speaker 1>a former editor at Pop Sugar, which had taken down

0:12:25.200 --> 0:12:28.520
<v Speaker 1>a number of articles that she'd written. So I went

0:12:28.559 --> 0:12:30.520
<v Speaker 1>online and it said that was taken down because there

0:12:30.559 --> 0:12:33.840
<v Speaker 1>inconsistencies or didn't live up to Pop Sugar standards. The

0:12:34.000 --> 0:12:37.000
<v Speaker 1>editor said that there were plagiarism issues with the stories.

0:12:37.880 --> 0:12:41.080
<v Speaker 2>So walk me through this, Like, what were you feeling

0:12:41.240 --> 0:12:42.000
<v Speaker 2>at this time?

0:12:42.840 --> 0:12:45.280
<v Speaker 1>I felt embarrassed to begin with. I think the pitch

0:12:45.280 --> 0:12:48.160
<v Speaker 1>call was I said something about like the ways that

0:12:48.679 --> 0:12:51.720
<v Speaker 1>health and money collide and twenty twenty five and Ontario

0:12:51.800 --> 0:12:53.920
<v Speaker 1>something like that, and it felt to me like the

0:12:53.920 --> 0:12:57.520
<v Speaker 1>pitches were some large language model taken that little prompt

0:12:57.600 --> 0:12:59.880
<v Speaker 1>and sending me back like what I wanted to hear.

0:13:00.040 --> 0:13:01.640
<v Speaker 1>And of course I was like, oh, yeah, I'm into this,

0:13:01.640 --> 0:13:03.559
<v Speaker 1>this makes sense. This is all about how health and

0:13:03.600 --> 0:13:06.280
<v Speaker 1>money collide in twenty twenty five, Like I was. Yeah,

0:13:06.320 --> 0:13:08.840
<v Speaker 1>I was embarrassed at being kind of like sent back

0:13:08.920 --> 0:13:11.760
<v Speaker 1>my own prompt in a way that felt like it

0:13:11.800 --> 0:13:14.079
<v Speaker 1>was just designed to appeal to me. And I got

0:13:14.080 --> 0:13:16.839
<v Speaker 1>fooled right as I moved through it, though, like I

0:13:16.880 --> 0:13:20.360
<v Speaker 1>don't know, I began to feel maybe it sounds overblown,

0:13:20.400 --> 0:13:22.840
<v Speaker 1>but like a sense of despair, Like at a certain point,

0:13:22.920 --> 0:13:25.560
<v Speaker 1>I'm tracking down all this stuff she's telling lies all

0:13:25.600 --> 0:13:27.840
<v Speaker 1>over the place. I go back to my inbox, which

0:13:27.840 --> 0:13:30.120
<v Speaker 1>still has a million pitches, and I'm going through it

0:13:30.160 --> 0:13:33.280
<v Speaker 1>and I'm seeing the ones that look so synthetic. So

0:13:33.360 --> 0:13:36.280
<v Speaker 1>AI and I started idly googling some of those authors,

0:13:36.280 --> 0:13:39.360
<v Speaker 1>and I see that they have places across the internet

0:13:39.360 --> 0:13:41.880
<v Speaker 1>as well. And I'm not going to do a deep

0:13:41.920 --> 0:13:44.640
<v Speaker 1>dive on every single freelancer that ends up in my inbox,

0:13:44.720 --> 0:13:46.960
<v Speaker 1>but like, how much of this stuff is out there?

0:13:47.120 --> 0:13:49.160
<v Speaker 1>Like how much of I don't know?

0:13:49.200 --> 0:13:49.640
<v Speaker 3>You know that?

0:13:50.040 --> 0:13:51.959
<v Speaker 1>That was depressing to me. I spent my whole life

0:13:51.960 --> 0:13:55.520
<v Speaker 1>writing journalism, and this feeling of this just being the

0:13:55.559 --> 0:13:59.880
<v Speaker 1>tip of a pretty like awful iceberg was dispiriting.

0:14:00.080 --> 0:14:00.360
<v Speaker 3>Guess.

0:14:01.280 --> 0:14:05.520
<v Speaker 2>After investigating all of this past work, Nick needed some answers.

0:14:06.040 --> 0:14:10.520
<v Speaker 2>He finally decided that he was ready to confront Victoria.

0:14:10.600 --> 0:14:13.000
<v Speaker 1>I wanted to talk to her. I had like so

0:14:13.120 --> 0:14:17.640
<v Speaker 1>many questions about who this person really is, Like I

0:14:17.679 --> 0:14:20.680
<v Speaker 1>as I'm in this world, I'm reading their writing right,

0:14:20.720 --> 0:14:22.920
<v Speaker 1>and I'm and some of its first person stuff, and

0:14:23.080 --> 0:14:26.000
<v Speaker 1>it's all obviously all a mix of true and false

0:14:26.240 --> 0:14:28.760
<v Speaker 1>or completely false. I'm trying to figure out where they are.

0:14:28.920 --> 0:14:31.120
<v Speaker 1>It feels like from some other earlier writing and some

0:14:31.160 --> 0:14:34.040
<v Speaker 1>other social media that they're they may be based in Nigeria,

0:14:34.120 --> 0:14:36.760
<v Speaker 1>or at least or from Nigeria. I said, a ton

0:14:36.760 --> 0:14:39.160
<v Speaker 1>of questions for this person, so I wanted to get

0:14:39.200 --> 0:14:40.800
<v Speaker 1>them on the phone. So I so I did that,

0:14:40.840 --> 0:14:43.120
<v Speaker 1>emailed and said, let's let's chat about this piece.

0:14:43.480 --> 0:14:46.480
<v Speaker 2>Nick emails Victoria asking if he can talk more about

0:14:46.520 --> 0:14:50.360
<v Speaker 2>her article. From a freelancers standpoint, and initial check in

0:14:50.440 --> 0:14:53.600
<v Speaker 2>call is a pretty normal next step toward officially being

0:14:53.640 --> 0:14:58.080
<v Speaker 2>accepted and being paid. So Victoria agreed to a video

0:14:58.160 --> 0:15:00.280
<v Speaker 2>call and they set it up for later that week.

0:15:02.400 --> 0:15:04.760
<v Speaker 1>In the ten minutes before that was about to start,

0:15:04.960 --> 0:15:07.720
<v Speaker 1>they emailed and said, I think I have internet connections.

0:15:07.760 --> 0:15:10.280
<v Speaker 1>Let's do this over audio. And then I got them

0:15:10.280 --> 0:15:12.720
<v Speaker 1>over the phone. And it was a strange conversation. It

0:15:12.760 --> 0:15:14.800
<v Speaker 1>was a it was a kind of surreal experience.

0:15:15.280 --> 0:15:15.920
<v Speaker 3>What happened.

0:15:16.560 --> 0:15:20.160
<v Speaker 1>So I guess in the fantasy version of this conversation

0:15:20.200 --> 0:15:22.640
<v Speaker 1>that I had been trying to plan out was I

0:15:22.680 --> 0:15:26.000
<v Speaker 1>would present them with some inconsistencies and they get bigger

0:15:26.040 --> 0:15:27.720
<v Speaker 1>and bigger, and a certain point they would just have

0:15:27.760 --> 0:15:30.880
<v Speaker 1>to admit these lies and then maybe we could actually

0:15:30.880 --> 0:15:32.280
<v Speaker 1>talk about who they were, and you know, were they

0:15:32.280 --> 0:15:34.880
<v Speaker 1>are real journalists who had gotten in over their heads.

0:15:34.880 --> 0:15:37.240
<v Speaker 1>Were they a scammer who had just like found some

0:15:37.280 --> 0:15:40.920
<v Speaker 1>easy marks with like overworked editors. That was the like

0:15:41.240 --> 0:15:44.000
<v Speaker 1>the unrealistic dream version of this conversation. How it actually

0:15:44.040 --> 0:15:47.000
<v Speaker 1>happened was I got on the phone someone very chipper

0:15:47.040 --> 0:15:49.520
<v Speaker 1>who send it to my ear like a young woman

0:15:49.560 --> 0:15:52.160
<v Speaker 1>with an African accent was on the phone. I asked

0:15:52.200 --> 0:15:54.600
<v Speaker 1>where she was based in Toronto, and she cheerfully said Bluer,

0:15:55.000 --> 0:15:56.840
<v Speaker 1>which is like if you said, whereabouts do you live

0:15:56.880 --> 0:16:00.440
<v Speaker 1>in New York and they just said Broadway right, okay,

0:16:00.800 --> 0:16:03.320
<v Speaker 1>But you know, this person had had googled the street

0:16:03.320 --> 0:16:05.520
<v Speaker 1>in Toronto, so I was I was already, you know,

0:16:05.760 --> 0:16:08.560
<v Speaker 1>well done. Yeah, And then I kind of just walked

0:16:08.560 --> 0:16:11.200
<v Speaker 1>them through some of the inconsistencies. And I didn't want

0:16:11.240 --> 0:16:13.360
<v Speaker 1>to spook them necessarily. I didn't want them to hang

0:16:13.440 --> 0:16:15.200
<v Speaker 1>up right away. So I asked them if they'd spoken

0:16:15.240 --> 0:16:17.520
<v Speaker 1>to the doctor. In her pitch, she said yes, she had.

0:16:17.720 --> 0:16:19.320
<v Speaker 1>I mentioned that, you know, we actually spoke to that

0:16:19.360 --> 0:16:21.800
<v Speaker 1>doctor and she says, she didn't speak to you, and

0:16:21.840 --> 0:16:23.800
<v Speaker 1>she said, oh, yes, of course, my personal assistant I

0:16:23.800 --> 0:16:27.080
<v Speaker 1>actually spoke with her. I didn't press on the fact

0:16:27.080 --> 0:16:30.800
<v Speaker 1>that like freelance journalists don't have personal assistance. I kind of, right, yeah,

0:16:31.440 --> 0:16:33.760
<v Speaker 1>let that drift past. I asked why I couldn't find

0:16:33.760 --> 0:16:36.320
<v Speaker 1>the stories in the Walrus and the Globe and Mail,

0:16:36.640 --> 0:16:39.160
<v Speaker 1>and she said most of those were done online only.

0:16:39.280 --> 0:16:41.080
<v Speaker 1>I emailed the editors from the Globe and the Walrus

0:16:41.080 --> 0:16:42.920
<v Speaker 1>and they said they've never worked with her. I asked

0:16:42.920 --> 0:16:44.840
<v Speaker 1>if she just recently moved to Toronto, because it seemed

0:16:44.880 --> 0:16:46.880
<v Speaker 1>like she was writing first person stories from London and

0:16:47.200 --> 0:16:49.400
<v Speaker 1>other places, and she said, yeah, she just moved this year.

0:16:49.960 --> 0:16:51.840
<v Speaker 1>I asked her why the stories were gone from Pop

0:16:51.880 --> 0:16:54.280
<v Speaker 1>Sugar and she said, oh, the editor moved on from

0:16:54.320 --> 0:16:56.840
<v Speaker 1>Pop Sugar, so they took down all the articles that

0:16:56.840 --> 0:16:59.600
<v Speaker 1>they worked talking with me, which is not how is

0:16:59.600 --> 0:17:02.400
<v Speaker 1>not how that works. Every question she just sort of answered,

0:17:02.480 --> 0:17:05.840
<v Speaker 1>very cheerfully, very upbeat, didn't seem to be shaken by it.

0:17:06.000 --> 0:17:08.600
<v Speaker 1>She had quick responses to every I mean, implausible, but

0:17:09.200 --> 0:17:11.920
<v Speaker 1>quick and ready. It's almost impressive. And then I kind

0:17:11.920 --> 0:17:14.680
<v Speaker 1>of moved on to the bigger inconsistencies.

0:17:14.760 --> 0:17:14.879
<v Speaker 4>Right.

0:17:15.080 --> 0:17:18.119
<v Speaker 1>I started talking about that story in the Scottish Journal.

0:17:18.480 --> 0:17:21.600
<v Speaker 1>I said that she'd quoted this professor Elaine Sutherland. I said,

0:17:21.600 --> 0:17:23.919
<v Speaker 1>I'd actually spoken to Elaine Sutherland, and she said she

0:17:23.960 --> 0:17:27.200
<v Speaker 1>has never spoken to you. And I realized in the

0:17:27.240 --> 0:17:30.000
<v Speaker 1>pause afterwards that she that she was gone, that she'd

0:17:30.080 --> 0:17:35.119
<v Speaker 1>hung up. She's not responded to any email since. I've

0:17:35.200 --> 0:17:36.080
<v Speaker 1>not heard from her since.

0:17:37.640 --> 0:17:40.360
<v Speaker 2>And it looks like that last question did it for Victoria.

0:17:40.640 --> 0:17:44.240
<v Speaker 2>She knew the jig was up, but Nick still wanted

0:17:44.280 --> 0:17:49.360
<v Speaker 2>to find out who is Victoria Goldie. We'll get into

0:17:49.400 --> 0:18:06.720
<v Speaker 2>that after the break. After that phone call, Victoria quietly

0:18:06.760 --> 0:18:08.440
<v Speaker 2>removed herself from the Internet.

0:18:08.920 --> 0:18:11.199
<v Speaker 1>In the days that followed, the x account that had

0:18:11.240 --> 0:18:15.160
<v Speaker 1>been posting some stuff under her name disappeared. Her mockrack page,

0:18:15.160 --> 0:18:17.760
<v Speaker 1>which is kind of a list of all a journalists

0:18:17.960 --> 0:18:22.679
<v Speaker 1>stories that went private. Her personal website went down. She

0:18:22.760 --> 0:18:26.280
<v Speaker 1>kind of disappeared after that phone call, and I say,

0:18:26.359 --> 0:18:29.080
<v Speaker 1>she I at this point, I don't not know who

0:18:29.400 --> 0:18:33.720
<v Speaker 1>who like Victoria Goldie is. The evidence suggests that this

0:18:33.800 --> 0:18:37.159
<v Speaker 1>person is either from or still lives in Nigeria. I

0:18:37.200 --> 0:18:41.119
<v Speaker 1>don't think it's a whole farm of people working together

0:18:41.400 --> 0:18:44.480
<v Speaker 1>and sending out stuff under one byline, although it could

0:18:44.520 --> 0:18:46.679
<v Speaker 1>be that. Something that crossed my mind a lot for

0:18:46.720 --> 0:18:49.520
<v Speaker 1>a while, I thought maybe there's no singular person attached

0:18:49.520 --> 0:18:51.320
<v Speaker 1>to this name at all. Maybe it's just, you know,

0:18:51.480 --> 0:18:53.600
<v Speaker 1>one of a dozen bylines that people are using to

0:18:54.040 --> 0:18:56.640
<v Speaker 1>get paid for some writing. I don't think that's the case.

0:18:56.680 --> 0:18:59.440
<v Speaker 1>That one of the earliest stories I found under that byline,

0:18:59.520 --> 0:19:01.880
<v Speaker 1>it was a couple months before Chatgebet came out, which

0:19:01.920 --> 0:19:04.600
<v Speaker 1>is kind of the demarketing line I think, after which

0:19:04.680 --> 0:19:07.320
<v Speaker 1>like you can't trust that a sentence was written by

0:19:07.359 --> 0:19:09.920
<v Speaker 1>a human, And it was like it felt different. Those

0:19:09.960 --> 0:19:13.800
<v Speaker 1>early stories were there's some grammatical mistakes, they're clumsier, they're

0:19:13.840 --> 0:19:16.919
<v Speaker 1>less smooth, less slick, but they seem to represent like

0:19:17.040 --> 0:19:19.720
<v Speaker 1>a real person talking about real things.

0:19:20.320 --> 0:19:23.160
<v Speaker 2>One article from that pre chatgpt era is from May

0:19:23.160 --> 0:19:26.360
<v Speaker 2>of twenty twenty two. It's a piece that Victoria Goldie

0:19:26.400 --> 0:19:29.439
<v Speaker 2>wrote for a website called black Ballad with the title

0:19:29.720 --> 0:19:33.480
<v Speaker 2>how I'm learning to navigate toxic positivity on social media,

0:19:34.080 --> 0:19:37.479
<v Speaker 2>And there's this line where she writes, quote, there's this

0:19:37.600 --> 0:19:41.080
<v Speaker 2>immense pressure to be productive. Most people like me are

0:19:41.160 --> 0:19:43.680
<v Speaker 2>tired and trying to survive day to day.

0:19:44.520 --> 0:19:47.159
<v Speaker 1>So I think this is a real individual. This is

0:19:47.200 --> 0:19:50.240
<v Speaker 1>not their name necessarily, but I think this is one individual.

0:19:50.240 --> 0:19:52.320
<v Speaker 1>And you see a kind of continuity of interest throughout

0:19:52.320 --> 0:19:56.040
<v Speaker 1>their into key dramas, they're into afrobeats. There's some continuity

0:19:56.080 --> 0:19:58.800
<v Speaker 1>of which feels like a person there. I can't know

0:19:58.880 --> 0:20:01.040
<v Speaker 1>for sure, but I think this as an individual who,

0:20:01.119 --> 0:20:03.440
<v Speaker 1>at least in those early stories, is writing about going

0:20:03.480 --> 0:20:06.520
<v Speaker 1>online like the rest of us and making her miserable

0:20:06.800 --> 0:20:09.840
<v Speaker 1>seeing people who are presenting themselves as doing so well,

0:20:09.960 --> 0:20:12.920
<v Speaker 1>seeing all this hustle culture and they're just struggling to

0:20:12.920 --> 0:20:13.800
<v Speaker 1>get by a day to day.

0:20:15.160 --> 0:20:18.760
<v Speaker 2>So in the beginning she was writing about going online

0:20:18.760 --> 0:20:22.520
<v Speaker 2>seeing hustle culture, getting kind of depressed. And that's what

0:20:22.560 --> 0:20:25.600
<v Speaker 2>the hustle culture Rose is selling you right now, is hey,

0:20:25.720 --> 0:20:27.639
<v Speaker 2>here's how to make a bunch of money by cranking

0:20:27.680 --> 0:20:32.360
<v Speaker 2>out content with AI. So kind of can't beat them,

0:20:32.800 --> 0:20:33.720
<v Speaker 2>join them type thing.

0:20:34.359 --> 0:20:36.760
<v Speaker 1>Yeah, So in the end, like I don't know who

0:20:36.760 --> 0:20:40.080
<v Speaker 1>this person is, I have some sympathy for them.

0:20:41.040 --> 0:20:43.639
<v Speaker 2>After all of this, Nick decided to make his investigation

0:20:43.760 --> 0:20:46.359
<v Speaker 2>public and he published a piece in the local and

0:20:46.520 --> 0:20:49.080
<v Speaker 2>that's actually where I first read about it. The title

0:20:49.119 --> 0:20:52.800
<v Speaker 2>of his piece was called Investigating a Possible Scammer in

0:20:52.880 --> 0:20:54.280
<v Speaker 2>Journalism's AI era.

0:20:54.960 --> 0:20:57.679
<v Speaker 1>We published the piece, it got a crazy response, but

0:20:57.720 --> 0:21:00.400
<v Speaker 1>I think the most like telling response. All I got

0:21:00.480 --> 0:21:03.040
<v Speaker 1>was so many emails from editors from around the world,

0:21:03.119 --> 0:21:05.720
<v Speaker 1>which was kind of mind blowing. Victoria Goldie had been

0:21:05.760 --> 0:21:11.600
<v Speaker 1>pitching like everywhere, everywhere, from nature to an Oakland startup

0:21:11.640 --> 0:21:14.520
<v Speaker 1>that began like a few weeks ago doing book reviews.

0:21:14.960 --> 0:21:17.719
<v Speaker 1>Someone on the West Coast who was running this tiny

0:21:17.920 --> 0:21:21.119
<v Speaker 1>indigenous run publication said that they had just assigned a

0:21:21.160 --> 0:21:23.080
<v Speaker 1>story to her and they just, oh my god, did

0:21:23.119 --> 0:21:24.600
<v Speaker 1>a quick Google to her and they're like, thank you

0:21:24.640 --> 0:21:27.200
<v Speaker 1>so much. We were very close to sending the contract.

0:21:27.680 --> 0:21:30.879
<v Speaker 1>Someone else from this publication called Republic. They phoned me

0:21:30.960 --> 0:21:32.679
<v Speaker 1>up there in the middle of a second draft with her.

0:21:32.880 --> 0:21:35.960
<v Speaker 1>They were editing a story about people living in their

0:21:36.080 --> 0:21:39.520
<v Speaker 1>vans in the American Southwest that was based on Victoria's

0:21:39.560 --> 0:21:41.760
<v Speaker 1>own experience living in a van in the American Southwest.

0:21:41.840 --> 0:21:46.119
<v Speaker 1>The Financial Times like around the world. This this person

0:21:46.320 --> 0:21:49.240
<v Speaker 1>was really incredibly prolific.

0:21:51.520 --> 0:21:54.720
<v Speaker 2>Just the sheer volume of pitches that Victoria Goldie was

0:21:54.720 --> 0:21:57.320
<v Speaker 2>sending out might give us a hint at her motivations.

0:21:57.760 --> 0:22:01.600
<v Speaker 2>Journalism is famously not a very lucrative career, especially for

0:22:01.640 --> 0:22:05.200
<v Speaker 2>a freelance journalist. Ask me how I know, and maybe

0:22:05.400 --> 0:22:09.520
<v Speaker 2>pre ai Victoria was having trouble finding work, But if

0:22:09.520 --> 0:22:12.359
<v Speaker 2>you can start writing for dozens of places at once,

0:22:12.960 --> 0:22:14.159
<v Speaker 2>that might start to add up.

0:22:14.920 --> 0:22:16.560
<v Speaker 1>I guess part of what's interesting to me is, I

0:22:16.560 --> 0:22:19.480
<v Speaker 1>think we were offering two thousand bucks for this story.

0:22:19.520 --> 0:22:22.280
<v Speaker 1>That's our rate Dwell paid, you know, I think she

0:22:22.359 --> 0:22:24.680
<v Speaker 1>got fifteen hundred dollars for that or something in that

0:22:24.760 --> 0:22:26.960
<v Speaker 1>Dwell article. If you actually had interviewed ten of the

0:22:27.000 --> 0:22:29.720
<v Speaker 1>top designers in the world, like you know that fifteen

0:22:29.760 --> 0:22:32.040
<v Speaker 1>hundred dollars, it doesn't go far. If you entered into

0:22:32.119 --> 0:22:34.480
<v Speaker 1>chat GPT and get this thing out of five minutes later,

0:22:34.520 --> 0:22:36.600
<v Speaker 1>that's a pretty good return, right.

0:22:37.440 --> 0:22:39.640
<v Speaker 2>And for a long time she got away with it.

0:22:40.000 --> 0:22:43.760
<v Speaker 2>Victoria Goldie had dozens of articles across the Internet for

0:22:43.920 --> 0:22:47.119
<v Speaker 2>a bunch of really well respected publications. By the end

0:22:47.119 --> 0:22:50.760
<v Speaker 2>of Nick's investigation, four of her old articles had come down,

0:22:51.080 --> 0:22:53.480
<v Speaker 2>but as of this recording, there are still a few

0:22:53.520 --> 0:22:59.880
<v Speaker 2>out there. How do we change editorial processes? I mean,

0:23:00.040 --> 0:23:02.760
<v Speaker 2>how is the local change their editorial process or they're

0:23:02.760 --> 0:23:06.359
<v Speaker 2>different hoops that somebody has to jump through to submitted article.

0:23:06.560 --> 0:23:09.920
<v Speaker 1>I mean, yeah, you got a big response on the internet,

0:23:09.960 --> 0:23:11.840
<v Speaker 1>but it got a big response in our newsroom. Like

0:23:11.880 --> 0:23:13.760
<v Speaker 1>we've been thinking about how do we do what we

0:23:13.800 --> 0:23:16.320
<v Speaker 1>do going forward in this era? Like I said at

0:23:16.320 --> 0:23:18.080
<v Speaker 1>the beginning, we want to work with new writers, right,

0:23:18.119 --> 0:23:20.000
<v Speaker 1>like this is our whole thing. I owe my career to.

0:23:20.119 --> 0:23:22.800
<v Speaker 1>Like a pitch used to be this key that could

0:23:22.800 --> 0:23:24.479
<v Speaker 1>open the door even if you didn't know anybody at

0:23:24.480 --> 0:23:26.960
<v Speaker 1>a publication. If you could write an amazing pitch, that's

0:23:27.000 --> 0:23:29.840
<v Speaker 1>a way into places. A pitch now is not connected

0:23:30.080 --> 0:23:32.400
<v Speaker 1>or is not necessarily connected to any human being. Right,

0:23:32.960 --> 0:23:35.520
<v Speaker 1>So when we are trying to find new writers, what

0:23:35.520 --> 0:23:38.480
<v Speaker 1>do you do? It's made our jobs difficult. Like we're

0:23:38.480 --> 0:23:41.040
<v Speaker 1>talking about I'm going to do more coffees with young freelancers.

0:23:41.040 --> 0:23:43.359
<v Speaker 1>I'm going to phone their editors and but this is

0:23:43.400 --> 0:23:45.760
<v Speaker 1>all like, this is all a bummer. This all takes

0:23:45.800 --> 0:23:48.919
<v Speaker 1>so much time, it takes so much money. Everyone's already overworked.

0:23:49.080 --> 0:23:51.200
<v Speaker 1>And I think the fear is that you're just gonna

0:23:51.200 --> 0:23:52.760
<v Speaker 1>work with people you already know. You're just going to

0:23:52.880 --> 0:23:55.919
<v Speaker 1>work with established people. You're never going to take a

0:23:56.000 --> 0:23:58.520
<v Speaker 1>chance on someone new because did they write their pitch?

0:23:58.800 --> 0:23:59.960
<v Speaker 1>Did they write their previous story?

0:24:00.359 --> 0:24:01.719
<v Speaker 3>Like you can't trust anything?

0:24:01.920 --> 0:24:07.040
<v Speaker 1>Yeah, yeah, So I am doing more phone calls earlier

0:24:07.080 --> 0:24:10.240
<v Speaker 1>in the process. With writers, at least you get a

0:24:10.280 --> 0:24:12.840
<v Speaker 1>sense of a person. You know, you're talking to a

0:24:12.880 --> 0:24:15.960
<v Speaker 1>real person. We're like, we've beefed up our fact checking process,

0:24:16.160 --> 0:24:19.680
<v Speaker 1>so people send annotated drafts or checking through them. I

0:24:19.720 --> 0:24:22.200
<v Speaker 1>think if you had robust fact checking, like this story

0:24:22.240 --> 0:24:25.639
<v Speaker 1>is not getting into twelve magazine, right. The depressing thing

0:24:25.680 --> 0:24:29.399
<v Speaker 1>about this scammer as opposed to like classic scammers like

0:24:29.800 --> 0:24:32.680
<v Speaker 1>Jason Blair or like Stephen Glass, Right, like people making

0:24:32.760 --> 0:24:34.480
<v Speaker 1>up stuff in the past. They were doing that in

0:24:34.520 --> 0:24:36.960
<v Speaker 1>a world or journalism has a certain amount of power

0:24:37.000 --> 0:24:38.720
<v Speaker 1>and prestige, and they're doing it for that reason. This

0:24:38.880 --> 0:24:42.480
<v Speaker 1>is people who are taking advantage of an ecosystem that

0:24:42.600 --> 0:24:47.200
<v Speaker 1>is like already pretty broken and yeah and overworked, overworked. Yeah,

0:24:47.200 --> 0:24:49.840
<v Speaker 1>there's not fact checkers. I don't think an editor at

0:24:49.840 --> 0:24:52.720
<v Speaker 1>some of these places could have thought twice about some

0:24:52.840 --> 0:24:55.800
<v Speaker 1>of this stuff. Even so, Yeah, to combat it, it

0:24:55.840 --> 0:24:58.080
<v Speaker 1>would take more money, would take more resources, it would

0:24:58.080 --> 0:25:00.760
<v Speaker 1>take journalists spending more time in that work. And we're

0:25:00.800 --> 0:25:03.159
<v Speaker 1>already at a point where like we're going in the

0:25:03.200 --> 0:25:04.200
<v Speaker 1>opposite direction.

0:25:04.040 --> 0:25:07.280
<v Speaker 2>Right, Yeah, I mean when when you said robust fact checking,

0:25:07.480 --> 0:25:10.440
<v Speaker 2>any journalists listening to this are probably laughing and or

0:25:10.480 --> 0:25:13.440
<v Speaker 2>spit out their coffee or crying into their coffee, probably

0:25:14.119 --> 0:25:16.600
<v Speaker 2>because that's one of the things that gets.

0:25:16.320 --> 0:25:17.880
<v Speaker 3>Cut, you know.

0:25:17.920 --> 0:25:22.080
<v Speaker 2>That's like do organizations have fact checkers, Yes, but maybe

0:25:22.080 --> 0:25:24.760
<v Speaker 2>they used to have five and they have one or

0:25:25.520 --> 0:25:27.840
<v Speaker 2>the editor maybe is being asked to handle all that

0:25:27.920 --> 0:25:31.320
<v Speaker 2>work and so we got to spend money to clean

0:25:31.359 --> 0:25:34.399
<v Speaker 2>it up. Yeah, And I mean this is it's interesting

0:25:34.440 --> 0:25:37.600
<v Speaker 2>because this isn't really necessarily about AI. You know, I

0:25:37.960 --> 0:25:41.560
<v Speaker 2>hate to echo the AI boosters here, right, but this

0:25:41.600 --> 0:25:44.600
<v Speaker 2>is something that's often said, AI is a tool, and

0:25:44.640 --> 0:25:48.439
<v Speaker 2>this is somebody who and maybe multiple somebodies who are

0:25:48.520 --> 0:25:51.640
<v Speaker 2>using just a tool that is more efficient.

0:25:51.720 --> 0:25:54.480
<v Speaker 1>Exactly the scale would be impossible with it at the

0:25:54.480 --> 0:25:57.639
<v Speaker 1>efficiency I would not have heard from dozens of editors

0:25:57.680 --> 0:26:00.720
<v Speaker 1>around the world had been pitched stories that fit pretty

0:26:00.760 --> 0:26:03.920
<v Speaker 1>well into their niche Like we're kind of we're way off.

0:26:04.280 --> 0:26:06.520
<v Speaker 1>That's only possible with the technology, right, But yeah, this

0:26:07.000 --> 0:26:09.760
<v Speaker 1>is this is human beings using that technology to do

0:26:09.840 --> 0:26:11.359
<v Speaker 1>something that they've done for a while.

0:26:12.920 --> 0:26:15.600
<v Speaker 2>But there is something new that comes with these AI

0:26:15.720 --> 0:26:19.960
<v Speaker 2>models that have scraped everything ever written. It's changing people's

0:26:20.000 --> 0:26:23.879
<v Speaker 2>definitions of truth and lies in ways that journalists haven't

0:26:23.880 --> 0:26:26.200
<v Speaker 2>had to think about before. And I'm not just saying,

0:26:26.240 --> 0:26:29.080
<v Speaker 2>you know, oh truth, there's important, lies or bad. I

0:26:29.200 --> 0:26:31.439
<v Speaker 2>just mean that it's getting weird.

0:26:33.320 --> 0:26:34.960
<v Speaker 1>So one of the responses I got from one of

0:26:35.000 --> 0:26:38.200
<v Speaker 1>these experts who I followed up on, who had quotes

0:26:38.240 --> 0:26:40.480
<v Speaker 1>made up by Victoria Goldie for one of these stories,

0:26:40.720 --> 0:26:43.040
<v Speaker 1>they said, that's not me. I don't remember speaking to

0:26:43.080 --> 0:26:45.840
<v Speaker 1>this writer, but that sounds like something I would say,

0:26:45.880 --> 0:26:48.199
<v Speaker 1>And I'm okay with that material being out there, and

0:26:48.240 --> 0:26:53.120
<v Speaker 1>that like wow, that took me back, Like that that

0:26:53.200 --> 0:26:55.760
<v Speaker 1>indicates that we are part of a world where, like

0:26:56.240 --> 0:26:57.640
<v Speaker 1>I think we already know that a lot of people

0:26:57.680 --> 0:27:00.280
<v Speaker 1>don't mind the slap, like the slop is ever where,

0:27:00.359 --> 0:27:03.600
<v Speaker 1>because people are fine with it if it's that's up

0:27:04.440 --> 0:27:08.280
<v Speaker 1>some of their preconceived ideas or it seems close enough, right.

0:27:08.640 --> 0:27:13.080
<v Speaker 1>And the fact that individuals themselves are okay with being represented,

0:27:13.440 --> 0:27:15.359
<v Speaker 1>you know it's close enough to what they said or

0:27:15.400 --> 0:27:17.639
<v Speaker 1>what they believe, that makes me think like we are

0:27:17.680 --> 0:27:20.520
<v Speaker 1>in a we're entering a whole new paradigm about like

0:27:20.560 --> 0:27:21.960
<v Speaker 1>what truth even is.

0:27:22.040 --> 0:27:28.600
<v Speaker 2>And that is like I don't know, all right, counterpoint,

0:27:28.800 --> 0:27:33.240
<v Speaker 2>what if it doesn't matter that we've got AI journalists. Look,

0:27:33.520 --> 0:27:37.360
<v Speaker 2>you wanted an article on the kind of creeping privatization

0:27:38.000 --> 0:27:42.439
<v Speaker 2>of the healthcare system. You have a general idea of

0:27:42.440 --> 0:27:44.760
<v Speaker 2>what some of these people are going to say, and

0:27:45.320 --> 0:27:48.400
<v Speaker 2>for a lay reader who isn't really familiar with everything,

0:27:48.800 --> 0:27:50.760
<v Speaker 2>they need to be brought up at speed on the basics. Right,

0:27:51.080 --> 0:27:54.360
<v Speaker 2>So what's the harm in having, you know, some chatbot

0:27:54.880 --> 0:27:57.240
<v Speaker 2>fill in the blanks for them and say some things

0:27:57.240 --> 0:28:00.760
<v Speaker 2>that are pretty plausible from a few experts. The AI

0:28:00.840 --> 0:28:03.760
<v Speaker 2>models is trained on everything they've said. It knows about

0:28:03.760 --> 0:28:05.200
<v Speaker 2>what they're gonna say.

0:28:05.320 --> 0:28:07.800
<v Speaker 1>Yeah, I mean, this is the argument that kind of

0:28:07.880 --> 0:28:10.120
<v Speaker 1>terrifies me, right, Like, I think, to be.

0:28:10.080 --> 0:28:11.399
<v Speaker 3>Clear, I'm not advocating for this.

0:28:11.480 --> 0:28:12.800
<v Speaker 1>No, no, no, I know, I know, I know, I

0:28:13.080 --> 0:28:17.440
<v Speaker 1>hear you taking the point of view, Like, I mean,

0:28:17.480 --> 0:28:19.679
<v Speaker 1>first of all, they don't know what individuals say, right,

0:28:19.720 --> 0:28:22.320
<v Speaker 1>everyone is surprising, and a I can only say what's

0:28:22.640 --> 0:28:26.480
<v Speaker 1>been said already. It's incapable of actually advancing a story.

0:28:26.480 --> 0:28:28.600
<v Speaker 1>And the news is about saying what is new, right,

0:28:28.640 --> 0:28:30.960
<v Speaker 1>and that is something that I think cannot be replaced

0:28:30.960 --> 0:28:34.159
<v Speaker 1>at all. But more fundamentally, like I don't know, we

0:28:34.400 --> 0:28:36.359
<v Speaker 1>just can't be cool with having stuff that is not

0:28:36.400 --> 0:28:39.120
<v Speaker 1>true presented as true because it's close enough, or it's

0:28:39.240 --> 0:28:42.960
<v Speaker 1>like we're losing that grip on you know, when someone's

0:28:43.040 --> 0:28:44.960
<v Speaker 1>quoted they have to have said a thing, when when

0:28:45.160 --> 0:28:47.360
<v Speaker 1>like I feel like I'm losing my mind as we're

0:28:47.960 --> 0:28:51.720
<v Speaker 1>kind of moving away from actually recognizing that, like when

0:28:51.720 --> 0:28:53.240
<v Speaker 1>you say stuff's true, it's got to be It's got

0:28:53.280 --> 0:28:54.920
<v Speaker 1>to be true. Otherwise, what are we even? How can

0:28:54.960 --> 0:28:56.920
<v Speaker 1>we talk? How can we argue? How can we agree

0:28:56.960 --> 0:28:57.440
<v Speaker 1>on anything?

0:28:58.200 --> 0:29:00.960
<v Speaker 2>Yeah, I think we were, we were in the post

0:29:01.000 --> 0:29:03.400
<v Speaker 2>truth thing already. It's like post post truth.

0:29:03.640 --> 0:29:05.920
<v Speaker 1>Yeah, I think like university professors have been here for

0:29:05.920 --> 0:29:07.720
<v Speaker 1>a while already, Like it's been a few years of

0:29:08.000 --> 0:29:10.360
<v Speaker 1>them dealing with this world. But for me personally, in

0:29:10.400 --> 0:29:13.520
<v Speaker 1>like my little corner of magazine journalism, I haven't been

0:29:13.520 --> 0:29:15.040
<v Speaker 1>confronted with it until now.

0:29:15.160 --> 0:29:16.120
<v Speaker 3>You know, it's interesting you.

0:29:16.040 --> 0:29:17.440
<v Speaker 2>Bring that up, because I think a lot of people

0:29:17.480 --> 0:29:21.760
<v Speaker 2>are curious about how university and professors or high school

0:29:21.800 --> 0:29:23.840
<v Speaker 2>professors or whatever dealing with the fact that, yeah, a

0:29:23.880 --> 0:29:26.280
<v Speaker 2>lot of their students are probably just typing a prompt

0:29:26.280 --> 0:29:28.880
<v Speaker 2>into CHATGBT and then you know, cranking out a thousand

0:29:28.920 --> 0:29:31.920
<v Speaker 2>word essay. But that's a very controlled environment. You know,

0:29:32.080 --> 0:29:35.560
<v Speaker 2>it's a fish bowl. If somebody at the university down

0:29:35.640 --> 0:29:39.600
<v Speaker 2>the road chat gbt generates their history exam, that doesn't

0:29:39.640 --> 0:29:45.440
<v Speaker 2>affect me personally. But if somebody is chatgebt generating an

0:29:45.600 --> 0:29:49.960
<v Speaker 2>article about what the LAPD is doing or not doing, that.

0:29:49.880 --> 0:29:50.600
<v Speaker 3>Could affect me.

0:29:51.240 --> 0:29:54.200
<v Speaker 2>And we're already a stage where people just straight up

0:29:54.240 --> 0:29:58.000
<v Speaker 2>don't trust journalism at all. And what do we do

0:29:58.080 --> 0:30:01.760
<v Speaker 2>if the if the writers are lying to the editors,

0:30:02.680 --> 0:30:03.959
<v Speaker 2>or the writers don't even exist.

0:30:05.040 --> 0:30:09.680
<v Speaker 1>Yeah, yeah, I have no answers. It's weighing on me.

0:30:12.480 --> 0:30:15.200
<v Speaker 2>As of right now, as I'm recording this, not all

0:30:15.240 --> 0:30:18.400
<v Speaker 2>of Victoria Goldie's work is gone. So among that work

0:30:18.440 --> 0:30:21.560
<v Speaker 2>are three articles that she wrote for Business Insider. Two

0:30:21.560 --> 0:30:24.320
<v Speaker 2>of those were published just before chat ChiPT dropped in

0:30:24.320 --> 0:30:26.720
<v Speaker 2>November of twenty twenty two, and one came out right

0:30:26.760 --> 0:30:30.280
<v Speaker 2>after I dug into all of these articles. Each one

0:30:30.360 --> 0:30:32.800
<v Speaker 2>is an interview with one person who, as far as

0:30:32.840 --> 0:30:35.680
<v Speaker 2>I can tell, seems to actually exists. The articles are

0:30:35.800 --> 0:30:39.240
<v Speaker 2>kind of boring, but they seem legit. But there's also

0:30:39.480 --> 0:30:42.200
<v Speaker 2>this really interesting article that she wrote for The Guardian,

0:30:42.280 --> 0:30:44.600
<v Speaker 2>and this one only came out in September of twenty

0:30:44.640 --> 0:30:48.440
<v Speaker 2>twenty five. So Victoria Goldie writes about hanging out in

0:30:48.520 --> 0:30:52.040
<v Speaker 2>East London and going to raves and soccer games, mixing

0:30:52.120 --> 0:30:56.080
<v Speaker 2>with different people, learning about different cultures, and near the

0:30:56.280 --> 0:31:00.680
<v Speaker 2>end of this really personal essay, she writes, quote, the

0:31:00.720 --> 0:31:04.920
<v Speaker 2>future of our music is not written by algorithm. That

0:31:05.120 --> 0:31:09.000
<v Speaker 2>article has been taken down? So was that one real?

0:31:09.480 --> 0:31:12.480
<v Speaker 2>Is any of this real? How many of us even care?

0:31:15.320 --> 0:31:15.680
<v Speaker 4>All right?

0:31:15.840 --> 0:31:17.960
<v Speaker 2>I don't want to lose sight of how all this started.

0:31:18.360 --> 0:31:21.400
<v Speaker 2>Nick has finally put his investigation to rest, and the

0:31:21.440 --> 0:31:24.840
<v Speaker 2>local special on healthcare in Canada is now out and

0:31:24.880 --> 0:31:26.760
<v Speaker 2>it's ready for you to read it. And if you're

0:31:26.760 --> 0:31:29.200
<v Speaker 2>wondering if they're going to put something about Victoria Goldie

0:31:29.200 --> 0:31:30.640
<v Speaker 2>in there, I.

0:31:30.560 --> 0:31:32.600
<v Speaker 1>Don't think so. I don't think so. We're moving moving on,

0:31:32.720 --> 0:31:33.360
<v Speaker 1>We're moving on.

0:31:33.680 --> 0:31:35.760
<v Speaker 3>Moving on, moving on. I dig it.

0:31:38.840 --> 0:31:41.640
<v Speaker 2>Thank you so much for checking out another episode of

0:31:41.720 --> 0:31:43.480
<v Speaker 2>kill Switch. If you want to talk to us, you

0:31:43.520 --> 0:31:47.120
<v Speaker 2>can email us at kill Switch at Kaleidoscope dot NYC

0:31:47.640 --> 0:31:48.680
<v Speaker 2>or on Instagram.

0:31:48.760 --> 0:31:50.360
<v Speaker 3>We're at kill Switch Pod.

0:31:50.760 --> 0:31:52.800
<v Speaker 2>And before you move on to doing something else, maybe

0:31:52.840 --> 0:31:55.000
<v Speaker 2>think of leaving the show or a review. It helps

0:31:55.040 --> 0:31:57.400
<v Speaker 2>other people find the show which in turn helps us

0:31:57.480 --> 0:32:00.360
<v Speaker 2>keep doing our thing. Kill Switch is hosted by Me

0:32:00.560 --> 0:32:04.520
<v Speaker 2>Dexter Thomas. It's produced by Shena Ozaki, Darluck Potts and

0:32:04.680 --> 0:32:08.080
<v Speaker 2>Julian Nutter. Our theme song is by Me and Kyle Murdoch,

0:32:08.240 --> 0:32:11.920
<v Speaker 2>and Kyle also mixes a show from Kaleidoscope. Our executive

0:32:11.960 --> 0:32:17.200
<v Speaker 2>producers are Ozma Lashin from Geshatgadour and Kate Osborne from iHeart.

0:32:17.240 --> 0:32:20.560
<v Speaker 2>Our executive producers of Katrina Norville and Nikki E Tour

0:32:21.200 --> 0:32:24.840
<v Speaker 2>catch on next week

0:32:30.760 --> 0:32:36.600
<v Speaker 4>Goodbye