E292 - Rowenna Fielding, Director, Miss IG Geek (United Kingdom)
[00:00] Debbie Reynolds: The personal views expressed by our podcast guests are their own and are not legal advice or official statements by their organizations.
[00:11] Hello, my name is Debbie Reynolds. They call me the Data Diva. This is the Data Diva Talks privacy podcast where we discuss data privacy issues with industry leaders around the world with information that businesses need to know.
[00:24] Now,
[00:25] I have a very special guest on the show, Rowenna Fielding.
[00:30] She is the director of Ms. I.G.
[00:33] geek in England, correct?
[00:36] Rowenna Fielding: That's right, yep.
[00:38] Debbie Reynolds: Well, I've been a fan of you for so long.
[00:42] I've looked at your work over the years. I really love your comments.
[00:46] And, you know, I have to say that a lot of people that I really look up to and respect in the data protection,
[00:56] data privacy, data sphere,
[00:59] they really talk very highly of you. So you're always on everyone's list in terms of someone to think about, someone to look at. And I follow you for so many years.
[01:09] And so I saw your name pop up on LinkedIn. I said, oh God, why haven't I invited her on the show? So we're rectifying that right now. So welcome to the show.
[01:19] Rowenna Fielding: Awesome. I'm really pleased to have been invited. It's a great honor. Love your work.
[01:25] Debbie Reynolds: Yeah, well, I would love to know. You fascinate me because you're.
[01:30] I feel like you're a data person like me. So you think about data in a lot of different lenses. Right? So it's not just data protection,
[01:38] it's not just governance, it's not just ethics. It's kind of all these things put together and the way that you talk about them.
[01:45] People, people can tell or they will be able to tell. You have a very deep, deep knowledge and expertise in that area. But how did you become the IG geek?
[01:56] Rowenna Fielding: I never took my A levels and so I didn't go to university.
[02:00] I went to work in the West End as a spotlight operator, which was great fun, but had no future,
[02:07] no career advancement prospects or anything. So I just randomly decided to do some IT training because, you know, it was the big thing.
[02:15] Discovered my inner geek realized I had a kind of affinity for it.
[02:21] And then I went into IT security and then security management.
[02:26] It wasn't really fulfilling me.
[02:28] And then I got a job doing information governance and it turned out that most of what I was doing on a day to day basis involves data protection law.
[02:38] And I started looking into that and I got fascinated.
[02:42] And this is,
[02:44] this is more than 12 years ago now. Wow. Oh my goodness. I'm getting old.
[02:48] So, yeah, I kind of carried on on that path. Just sort of side slipped into doing data protection as a full time job and as a career. And then a couple of years back I sort of expanded that to include data ethics because ethics is a big topic at the moment.
[03:07] Our technological capabilities are outpacing our common sense quite fast. And so now I am a consultant. I work with organizations who want to do data protection or data information governance, data ethics.
[03:26] Not to tick boxes or because somebody's forcing them to, but because they genuinely don't want to cause harm to people.
[03:35] And so, yeah, I work with organizations that want to know how to do stuff right and how to think through stuff. And because I'm one of the world's biggest overthinkers, a lot of the questions I get asked are things that I've shower thoughts I've already had.
[03:52] So yeah, kind of help people by asking, what is it you're trying to achieve? Do you want to look good? Do you want to be able to dig yourself out of the holes that you fall into?
[04:01] Or do you want to do good,
[04:04] be good?
[04:05] Debbie Reynolds: Oh, that's great, that's excellent. I just want your scope or your thought on data ethics,
[04:13] your point of view there.
[04:15] So I feel as though people think,
[04:19] oh, we could just do all these technical things and we can pull up policies and then ethics somehow magically get addressed and it just doesn't happen that way. But I just want your thoughts on ethics and why it's important and how it connects to the work that you do.
[04:34] Rowenna Fielding: Yeah, absolutely. And you're right in what you say. People assume that because they are nice people trying to do beneficial things, that everything they do will therefore turn out to be nice and beneficial.
[04:48] But actually it's a lot easier and faster and cheaper to cause harm with data than it is to use it responsibly and safely.
[04:59] And I think a lot of that comes from people not really thinking about what data is in the first place.
[05:06] You've got this idea floating around that it's somehow facts or objective, neutral truth, when actually data is the output of social and technical processes created for a purpose by somebody with an agenda.
[05:21] So yeah, people think, oh, you know, there's all this data lying around in the wild, let's scoop it up and be useful with it. When actually what they're really doing is they're taking somebody else's ideas and observations and putting them to use without necessarily even thinking, is this appropriate data?
[05:40] Should I even have this data?
[05:42] So data ethics,
[05:43] it's a difficult topic to talk about because everybody wants to assume that they are just a nice person doing nice things. And if you start digging into actual outcomes and decision making frameworks and incentives and stuff, they feel like you're accusing them of being the bad guy.
[06:04] Even though what they may be doing may actually be indistinguishable from bad guy things.
[06:10] They're doing it by accident because they haven't thought about it.
[06:13] And ethics is a difficult one because when people say ethics, what they usually mean is cultural norms around social relations and altruism and,
[06:26] you know, protection of others.
[06:28] But ethics is just a, a code of values and beliefs.
[06:33] So there are people who genuinely believe that the right thing to do is to cause pain and distress to others in order to teach them the right way to do things.
[06:44] That's a code of ethics.
[06:46] It's not one that I personally think is great. But the first thing, what happens in data ethics is that there's a long protracted argument about the semantics of what ethics is and what it means.
[06:59] So that's difficult, but it's useful to frame it in terms of corporate social responsibility and environmental and social governance. That gives frameworks for, for what it is you're talking about.
[07:11] I like to get much more simple than that and kind of distill it down to don't be a git.
[07:17] Git, by the way, is Brit language for I suppose the American equivalent would be jerk.
[07:22] So yeah, don't be a jerk for those listeners across the pond.
[07:27] And that's basically what data protection and privacy law sort of boil down to, which is, yeah, don't be a, be careful about how you're treating other people or will happen to them because of what you're doing with their data.
[07:41] It's like health and safety for the intangible self.
[07:46] So health and safety for your non physical existence, although obviously it touches on the physical as well. So there's a lot of philosophy to get into when it comes to ethics.
[07:57] And yeah, most businesses don't want to sit, sit around and have a philosophy discussion. That sort of thing happens in the pub.
[08:04] So yeah, the first part of data ethic is, you know, asking like, who do you want to be? What do you want to achieve?
[08:11] What are your red lines? If child slavery were made legal tomorrow,
[08:16] would you go forth and adopt it on the basis that you have to remain competitive or, you know, if kicking puppies were mandated by the government,
[08:26] would you go and out and have your business k puppies in order to stay competitive? Like, what are your organization's values and standards basically?
[08:36] And that's a really hard question to answer because again, you know, people don't tend to think it through. They just assume that they're good people doing nice things.
[08:45] Debbie Reynolds: I think that's true. And then I think the thing that makes it even harder is that what we're talking about is harm that may not be tangible in the same way that we think of laws about harm.
[09:01] Right. So just like your example about someone kicking puppies. Right.
[09:06] I think that's kind of illegal in a lot of places, like harming animals and stuff like that.
[09:11] But if you do something that harms someone psychologically or, you know what I'm saying, harms them in some other way that's not physical, it's not something that you can see,
[09:23] I think that we have a hard time with that. That's where I feel ethics can play a really big part.
[09:29] Like, we have a lot of arguments in lawsuits in the US around privacy and data protection because people are like, oh, you use my data for this thing? And they're like, well, you weren't harmed,
[09:41] quote unquote. Like, no one stole your identity or no one did anything else. But then we have some laws, like we have a law in the US called the Biometric Information Privacy act,
[09:52] and,
[09:53] and a lot of lawyers in the US don't like that law because the harm is intangible. The harm is that you use someone's biometric in a way that they didn't want.
[10:03] And so I'm, I'm interested to see how all this plays out. But I think ethics plays like a huge part in trying to explain what that harm is.
[10:14] Rowenna Fielding: That's right. And actually the, the gdpr,
[10:17] either EU or UK version is actually quite helpful with that because it sets out the baseline, as in the fundamental rights and freedoms of the individual, as set out in the EU Charter of fundamental rights must be upheld.
[10:34] And there's about 50 of those. About a third of them relate to the governance of the EU itself. So they're not generally not applicable to most, most things. But yeah, that this law itself spells out what kind of the ethical objectives are, which is to uphold and not diminish or degrade or infringe or prevent people's fundamental human rights from being interfered with.
[11:02] So that makes life easier in some ways.
[11:07] In other ways it makes it difficult because data harms are stochastic effects. So usually in law you've got your sort of proximate cause did thing A, cause thing B.
[11:20] That is really hard to trace when you're talking about data harms because data is Infinitely replicable. You don't know how many people out there have got data about you.
[11:33] And it's a lot like ASEPS protocol in surgery.
[11:37] So when surgeons go into an operating theatre they are not trying to hunt down every individual germ by name and wipe it out in case that particular germ gets into a patient and causes an infection.
[11:52] Just make the whole thing as sterile as possible, prevent germs from getting to the patient.
[11:57] I think data ethics to me is the data world, the information age version of aseptic protocol. It's making sure that there are layers in place,
[12:10] systems, process,
[12:12] governance,
[12:13] culture,
[12:14] education,
[12:15] incentives very important that make using data in non harmful ways the default path, the path of, I mean it won't be least resistance, but the normatized approach.
[12:28] Unfortunately the compliance industry,
[12:32] the incentive there is to perform compliance not to necessarily be fulfilling the objectives stated in the law, which are, we've all seen what happens when people are allowed people's fundamental rights get tossed out the window or don't even exist.
[12:52] When it comes to data.
[12:53] Not a lot of people necessarily know this, but data protection law comes from human rights law, which comes from World War II and data protection law in particular because of how the, the Third Reich weaponized data against the people they were trying to get rid of with the enthusiastic collaboration of IBM at the time.
[13:16] So yeah, kind of going into the history of why this stuff is law can be helpful for helping people work with data ethics. But they've got to want to care.
[13:26] And that's getting harder and harder because data is everywhere.
[13:30] It's easy to generate, it's easy to use, it's easy to distribute. But thinking time is not in great supply and discussion time and room for people to have differing ideas and form consensus, it's just, it's, there's not room for it really in the modern world of business.
[13:53] So. Which makes it all the more important,
[13:55] Debbie Reynolds: I think, I think that's true.
[13:57] I want to talk a little bit with you about artificial intelligence. And the reason why I want to talk about this is that I very much like the EU's AI act and their focus on harm,
[14:12] where you know that concepts around data, this is hard in the US to be able to translate it that way because we don't think about law and regulation in that way.
[14:23] But the thing that's happening with AI, and I want your thoughts is that we have organizations now that are saying, oh I set up this chatbot, I set up this agent and it did this thing and I didn't do it like the AI did it.
[14:41] So, like, I'm not responsible or, or I didn't know that it was gonna harm this person in this way, but I'm not responsible. But what are your thoughts? I just feel like this is the new age, the dog ate my homework type of thing.
[14:53] Rowenna Fielding: So there is a quote from the American sociologist Edward Wilson,
[14:59] which goes something like, the problem humanity has is that we have Paleolithic emotions, medieval institutions, and godlike technology.
[15:09] And it is terrifically dangerous and it is the point of disaster.
[15:14] I feel that that could have. I mean, I think he said that in the 60s, but he, he could have said that about the AI. I'm doing air quotes around AI because AI is actually a marketing term.
[15:25] There is no single AI technology.
[15:27] There are a bunch of technologies which are labeled AI.
[15:32] So whether you're talking about machine learning,
[15:35] generating and working on statistical models for inference and prediction,
[15:40] for large language models, which are essentially also prediction about statistical word patterns,
[15:47] or the quant systems that are used in the financial world to try and read minds and predict the future,
[15:56] which has always been a thing humans are interested in and has always turned out to be impossible.
[16:02] Just takes a little while to come to terms with that fact.
[16:06] Yeah, there is no AI. What most people are talking about is either large language models which are presented as chatbots,
[16:15] or systems that purport to infer or predict behaviors or events.
[16:23] And I think if a machine learning model has been developed and designed by subject matter experts, that is experts in the thing they're applying it to, not just computer science experts,
[16:37] for a very narrow and specific use case,
[16:41] with very thorough training and development and review, feedback loops all the way through extremely careful deployment by,
[16:51] well,
[16:52] people who have been trained in the system's limitations and risks, not just its capabilities and ongoing monitoring with feedback loops for error correction,
[17:03] then it can be a wonderful, really useful thing.
[17:06] For example,
[17:07] the detection of tumours in X rays. Computers can see grayscale far better than humans.
[17:14] And that when done by medical experts and radiologists and stuff like that, you know, that's really successful. The problem comes when people use this technology to shortcut thinking, because there's really no substitute for.
[17:29] For thinking. And while one person can make a mistake,
[17:33] one person with an algorithm can propagate that mistake across the world to hundreds and thousands of places in the blink of an eye. And clearing up after it is a lot harder.
[17:48] So, yeah, I think that basically that this technology is a technology that humanity is simply not equipped to use safely or responsibly, but that's not actually going to make any difference, because it is burgeoning bubble.
[18:02] In fact, it's. It's kind of the new blockchain.
[18:05] You know, remember, blockchain was going to save the world and do all that, and now it's AI, but it's more dangerous because it's. The technology is presented as though it's as good as or better than a human.
[18:21] And that's not necessarily the case,
[18:24] but what it is good at is looking like it is as good as or better than a human until you start digging into it.
[18:31] So, actually, the Turing Test is the wrong way round.
[18:34] A Turing Test is a test of human gullibility.
[18:38] It's not a test of machine sentience. It's how easy is it to fool a human? And humans are really easy to fool and really hard to educate in the information age because our instincts are all just totally maladapted for a world that's based on the presumption of rationality and evidence and information and statistical probabilities that our squishy little brains just aren't equipped to handle.
[19:04] So, yeah, I think it's. I think it's the apocalypse, but it's a remarkably tedious one.
[19:09] Debbie Reynolds: I think so, too. Oh, my gosh. I was just laughing when you said about the Turing Test. And I've been saying this forever. I'm like, the Turing Test is about how easy people are to fool, not how smart computers are.
[19:21] Rowenna Fielding: Yeah.
[19:22] And that's without even getting into all the problematic questions about what is intelligence? I mean, it is impossible to have that conversation without at some point, somebody's starting, well, at some point, somebody bringing racism, sexism, classism, gender, cultural differences,
[19:41] physicalities. You know, ableism. Like, there is no definition of intelligence that doesn't just **** all over somebody.
[19:48] Debbie Reynolds: Well, yeah, right. Just like I would think about games like Trivial Pursuit or just education in general. Some educators,
[19:57] for them, education is learning some fact or some figure and being able to spout it. And to me, that's not learning,
[20:04] that's not knowledge. And so I feel like a lot of times when people say, oh, these AI tools, they have all this information. I'm like, but information isn't knowledge, and information is not wisdom.
[20:15] And so I think only humans can be wise. So even if you gave AI all the information in the world,
[20:21] you can't reason just like a human. And so one of my concerns is that we have tools, like you say in marketing speak, where we're saying, well,
[20:31] we can bypass the experts that know this thing, or we can bypass the humans and we can let the computer do these things. It's like, but that's not true and that's not real.
[20:42] And like we've never been able to do that. But the problem is that people think that it can. And to me that's like the danger.
[20:52] Right. So where you're giving a tool, like a headless tool,
[20:56] something that needs a human judgment and that gap is never going to be filled by technology.
[21:04] Rowenna Fielding: I mean the problem with filling that gap with technology is that the outcomes are become increasingly unpredictable and adverse.
[21:14] So I mean what we're seeing at the moment is people are filling that gap with technology because they think it's better or it's cheaper or it's definitely faster to make mistakes at scale than to have one person do something.
[21:27] Right.
[21:28] But the long term, and humans are not good with diffuse long term outcomes. It's basic gambling psychology as well. We would rather take a smaller win now and risk a bigger loss later than take a small loss now and bigger win later.
[21:44] And swipe gambling works. It's, it's so addictive. We're just not wired for the assumption that we're not wired to be able to keep to the assumptions and,
[21:56] you know, standards that we set for everybody else.
[22:00] Debbie Reynolds: Yeah.
[22:02] Rowenna Fielding: So yeah, it's great fun, isn't it?
[22:05] Debbie Reynolds: Yeah, it is. Absolutely. Well, what's happening in your world right now in technology or data that's concerning you most?
[22:13] Rowenna Fielding: Oh my goodness. Where to even start?
[22:17] Well, obviously there's the, the big data fairy, which is my version of the, the marketing term AI. And I call it that because people are to believe in this stuff like people are or were or have been willing to believe in magic.
[22:32] They don't understand how it works, but it looks like it does,
[22:36] and that's good enough. So the data fairy gets imbued with all these magical properties that it things it can supposedly do. Read minds, predict the future,
[22:46] change the world, make humans perfect,
[22:49] all that nonsense. And it's not the technology itself that scares me as much as the,
[22:55] the effect it has on people's critical thinking abilities.
[22:59] Because it's marketed to replace critical thinking cognitive labor. And there are already studies coming out showing that the less humans use their critical thinking facilities,
[23:11] the less able they are to use critical thinking. So actually, yes, AI makes you stupider to use it or lose it. So there's that and then there's kind of on the back of that there's the,
[23:23] the biometric inference. So the prediction and inference stuff, which is just like wholly rooted in phrenology and racism, ableism,
[23:34] sexism,
[23:35] the idea that there is a standard model human which coincidentally usually turns out to be a straight middle class white guy.
[23:44] So this standard human against whom everybody else is measured and found either suitably conforming or non conforming and that this standard human sort of sets the baseline for how all humans are supposed to be.
[23:59] And that's just not true.
[24:02] But even more than it being not true,
[24:05] it's actively harmful to people who especially whose physiology but also their neurological profile doesn't fit that template.
[24:15] So you've got that automated interviewing,
[24:18] sentiment analysis, behavioral prediction software that is resulting in neurodivergent people just not being able to get jobs because they can't get past the robot interviewer. You've got people with facial injuries, birthmarks, facial palsy, whatever,
[24:36] not being able to participate in the society because everything is facial recognition and the software doesn't even recognize them as human. Like can you imagine how soul destroying it would be to have everywhere you go, everything you try and do have these systems say no humans here, 404 human not found.
[24:56] Like that's appalling.
[24:58] And I mean, yeah, there's so many examples of why trying to read people's minds or figure out what they're going to do based on how they look. I mean it's stupid and it's dangerous, but it is incredibly lucrative because people want to believe it.
[25:15] So and then like tomorrow's big doom and gloom.
[25:19] And another thing is direct neural interface which has enormous potential to be horrifically dangerous,
[25:28] especially in the hands of people who see it as a commercial opportunity and not a horrifically dangerous risk.
[25:37] Debbie Reynolds: Absolutely.
[25:39] So I'm a person,
[25:41] I don't fit in any category,
[25:43] right?
[25:44] So a lot of the gaps that I see and I think one of the things that attracted me to data privacy and data protections because I was seeing the gaps.
[25:53] I live in the gap.
[25:54] So when I see people do these technologies, I'm like oh, like when you're talking about neurotechnology and brain and mind reading, I can guarantee you that you can't read my mind because you don't know my mind, right?
[26:07] You don't know me, so how can you read my mind, right? And then we have systems making assumptions about us. Like you're talking about the interviewing. Like, like I read a lot of those articles.
[26:18] Like oh, if someone looks up to the right, that means they're lying. It's like, but that doesn't, that's not true.
[26:26] Like you're acting like everyone's brain is the same or everyone's thinking is the same, also everyone's intention. Right. So everyone has different emotional cues, physical cues.
[26:38] And so I'm feeling like those things are being misread in these systems.
[26:43] But I love this conversation, by the way. I want to go to the philosophical plane with you for a minute. And so I. Another thing that concerns me about AI systems or the way these systems work is that the absence of data is also data,
[27:01] right? Yes, systems. And so people don't understand that. They think, oh yeah, we put it in and it gets this information.
[27:08] And so, you know, an example of this is that there was a famous company that was using artificial intelligence to look at resumes.
[27:21] Yeah, yes, yes.
[27:23] Rowenna Fielding: That AI came to the conclusion based on looking at who they had hired in the past, that they should only hire white guys called Jeff who played Cross in Ivy League college because they were hiring.
[27:38] Debbie Reynolds: Right,
[27:39] exactly, exactly. So they were filtering out anyone else who wasn't who didn't fit that profile. Right. But then a lot of the data of people who are different don't even get into those systems.
[27:52] But then you're trying to make a blanket assumption about everyone based on the limited data that you put into the system or the things that you decided are important or not.
[28:04] What are your thoughts?
[28:06] Rowenna Fielding: Yeah, absolutely. And that sort of goes back to what we were talking about at the start about understanding what data is and what data isn't. And it's not objective fact or truth or static.
[28:18] What it is is always incomplete that somebody has decided to cut out more else. That's all it can ever be. But yeah, so.
[28:28] And then systems are ending up creating self fulfilling prophecies. So for example, the, the machine learning models for offender sentencing.
[28:37] It turns out you are predicted to be more likely to reoffend if you are a person of colour.
[28:46] Not because people of colour are inherently criminal, but because of social history, social injustice, economic disparity and the fact that you're more likely to get caught and prosecuted as a person of color.
[29:02] So that is then creating a demographic disadvantage as a whole for that group of people and leading to that prophecy coming true because it's created the reality where it can only come true.
[29:16] I don't know if you've ever seen the movie Paycheck with Ben Affleck.
[29:21] So it's based on a Philip K. **** novel. It's kind of a cheesy movie, but quite good actually, if you like cheesy movies. So yeah, it kind of goes into the question of can the future be Predicted.
[29:30] Can it be altered a bit, like the whole minority report. Did you catch the ball? Yes, but it was going to fall. But how do you know? Because you caught it.
[29:40] Those are questions that they are philosophy questions and they're philosophy questions that in the main the computer science industry is disinterested in answering because they're big and scary and they're very diffuse and difficult to talk about.
[29:59] And you know, they don't win angel investments.
[30:02] So yeah, the self referential nature of these systems,
[30:08] they can only work with what's been put into them.
[30:11] What's been put into them is selected by somebody who has an objective to achieve and selected from what can be observed and generated and obtained and recorded.
[30:27] And because we're talking digital and not analog here, it's got to be normalized to an extent. It's got to have artificial boundaries imposed on it in order to be categorized into ones and zeros.
[30:40] Another example is the NYPD took a DNA sample from a crime scene and they went to a company that claimed to be able to create images of, you know, a person's likeness based on their DNA,
[30:56] which is highly dubious because I don't think they were factoring epigenetics into that, let alone environmental factors. And so, yeah, they got this image of who, of what this perpetrator was supposed to look like and then they fed it into fatal recognition software to find the person and of course ended up arresting a completely innocent person who also happened to be African American.
[31:19] Go figure. So it's just layer on layer of shaky assumptions and fudging and corner cutting and, you know, good enough for the moment. And, you know, when you stack all that up high enough,
[31:35] sooner or later it's going to topple over. And the people it lands on are usually the people that have least advantage in society.
[31:43] Debbie Reynolds: Wow, that's so true. That's so true. Thank you for doing this philosophical thing with me. I actually love it.
[31:49] One thing that concerns me now with like AI agents and the scary things we see people doing with technology,
[31:57] I feel like things are being created that even the makers are saying they have problems controlling.
[32:06] Right.
[32:07] And to me,
[32:08] and I could be going like way far left field, but to me that's what the nuclear age was about,
[32:18] which was we created something very dangerous and we have to find a way to control it and who controls it and for what purpose.
[32:28] So that's just my thought. What do you think?
[32:30] Rowenna Fielding: Yeah, I absolutely agree.
[32:32] I think that's humanity's relationship with technology since the dawn of time has been F A, F O.
[32:41] If you know, you know, we figure stuff out or we figure out how to use stuff and then a long time after we get around to figuring it out,
[32:51] how it works and why it works.
[32:54] Electricity was being used for things long before free electron theory and geometry was figured out, long before the underlying physics of it was figured out. I'm not a mathematician or a physicist, you can probably tell.
[33:10] I just know a bit about the history of science, technology.
[33:13] And so this is just a pattern of behavior that is repeated all the way throughout history.
[33:18] But the difference now is the scale of the technology and the impact and the consequences are bigger than they have ever been. I actually think that digital data technology is the most dangerous technology humanity has ever invented.
[33:37] It out dangers nuclear,
[33:40] nuclear power by an order of magnitude because the worst thing you can do to somebody with a blast of ionizing radiation is kill them.
[33:50] The worst thing you can do to somebody with their data is condemn them to a long life of misery,
[33:59] disadvantage,
[34:00] unfairness,
[34:03] radicalization,
[34:04] polarization, isolation,
[34:07] poor mental well being, poor physical well being.
[34:10] You know, it's kind of a life of misery is kind of worse in a lot of ways than a quick fiery ending. That's just my personal belief.
[34:20] I'm not saying that that's how everybody should think.
[34:24] And it's dangerous because it's so much more insidious.
[34:27] I mean, you can take a Geiger counter into a hot zone and figure out how much harmful radiation you're getting, which will give you, you know, a prognosis. But aside from a few academic and civil society institutes and higher education institutes around the world, nobody's really doing the same for data harms.
[34:50] There's the data justice lab, there's crack labs in Europe.
[34:54] There is work being done.
[34:56] But until the scale of harm reaches out and knocks the people at the top of the social order off their perches, there isn't a lot of incentive to do much about it.
[35:07] And again, that's how things go in society. Things change when the people with the most social capital and privilege start to be adversely affected by them. It was the same with food and drug safety,
[35:20] electrical supply safety, water safety,
[35:24] employment rights and labor laws, like everything automotive safety,
[35:29] yeah, there is has to be a critical mass of harm before the tide turns towards a pro safety culture.
[35:39] And the problem with data harms is identifying the source, the origins,
[35:45] in order to be able to say, well, this caused this, and then this happened. And this because it's so spread out and so diffuse,
[35:52] everybody can simply go, oh, we're just doing this little bit over here. It's, we didn't cause this. It's a lot like climate change.
[35:59] One person using loads of plastic bags,
[36:04] but 6 billion people wanting plastic bags, that's a different story.
[36:10] So yeah, again we've created a problem that is of a scale that's I think, outside our ability to manage. And it's all going to end in tears,
[36:19] but hopefully I won't be around to see it.
[36:24] That's my optimism of the day.
[36:26] Debbie Reynolds: Yeah, right, right, yeah. It's like a thousand paper cuts. Right? So every,
[36:31] so like every company is like, oh, well, I didn't do a whole part of this harm, but I did this one thing. So that one thing that I did to this person should not have harmed them.
[36:42] But a thousand other companies are doing the same thing. Right.
[36:45] So we're not thinking of it like a car. We're okay, the brakes went out on someone's car and they had an accident. We're saying just the force multiplier of the fact that these incremental harms can be happening to people.
[37:01] And it may not be evident in the moment, but over time it kind of snowballs into other things.
[37:07] Rowenna Fielding: Yeah, yeah, absolutely. And that's why I think the sort of asepsis protocol approach for data safety,
[37:16] I mean privacy, data protection, fundamental rights,
[37:20] they're all expressions of safety of the individual existing in society.
[37:26] So data safety is going to have to be about wall to wall preventative care,
[37:33] sterilizing the environment and the tools and the people and having reliable, consistent processes,
[37:44] just all that stuff.
[37:45] It's analogous, but this isn't one operating theater. This is an entire globe of different industries and businesses and cultures and nations and jurisdictions.
[37:59] And yeah, that's a,
[38:02] there's not really a way to get a handle on that without leaning towards authoritarianism,
[38:07] which is bad on its own.
[38:10] What do we do exactly?
[38:13] Debbie Reynolds: Exactly. Well, if power under the duvet is
[38:17] Rowenna Fielding: my answer,
[38:22] Debbie Reynolds: if it were the world according to you, and we did everything that you said, what will be your wish for data privacy, data protection,
[38:29] data anywhere in the world, whether that be human behavior,
[38:34] technology or regulation.
[38:37] Rowenna Fielding: Oh my goodness. Wow. Yeah, that's a big one. One wish.
[38:44] Debbie Reynolds: Oh, you can have more than one.
[38:47] Rowenna Fielding: Okay.
[38:48] So I mean, I would certainly want to see more funding,
[38:53] more prominence, more support,
[38:56] more engagement with and prioritization of those institutions that are doing data harm research and, you know, mapping things like, I forget, I think it was the Data Justice Lab, whoever it was, that figured out that certain demographics were not being shown, certain ads because it was assumed that they would be too poor and useless to take advantage of them.
[39:23] There's not a lot you can do without, you know, getting into, like, tyrannical dictator territory.
[39:29] But actually my wish would be that everybody just has and takes the time to think things through more than they are at the moment.
[39:42] I mean, in the modern workplace, thinking is not encouraged, it's not incentivized, it's not rewarded.
[39:48] Motion is action,
[39:51] whether it be meaningful and useful or not,
[39:54] because it can be measured.
[39:56] So, yeah, I would also like to get rid of the idea that data is inherently valuable because data is.
[40:05] Is always incomplete, it's always inadequate,
[40:11] it's always inaccurate and imprecise.
[40:15] It's the best we can do with what we have. But we've got to stop treating it as though it were any kind of gospel or God given pronouncement because it's just getting us into so much trouble.
[40:28] But then again, that requires thinking. So, yeah, time and inclination to think is my wish.
[40:34] Debbie Reynolds: I share your wish with you. Right. We do need time and inclination to think, so that's very wise words for you.
[40:43] Well, thank you so much. It's been such a pleasure to have you on the show.
[40:47] It's been many years in the making,
[40:50] so I'm happy to have you here and I love your thoughts. So thank you so much.
[40:54] Rowenna Fielding: I loved our chat very much. Thank you very much for inviting me to do it.
[41:00] Debbie Reynolds: This is great. This is great. Well, I'm sure we'll find ways we can collaborate together in the future.
[41:05] Rowenna Fielding: I would love that. And yeah, even just like a virtual meeting with a cup of tea to sort of geek out and rant at each other about the state of the world.
[41:15] Debbie Reynolds: Oh, totally. I love geeking out. I love geeking out. So that'll be great. We'll talk more philosophy. That'll be great.
[41:21] Yeah.
[41:25] All right, talk to you soon. Thank you so much.
[41:28] Rowenna Fielding: Take care and have an absolutely super day.
