Wellness, Play, and Professional Growth in the Digital Age with Joy Dean

Joy Dean, a commercial strategist and ad tech professional, shares her journey in the industry and her pioneering work with AI in eye tracking, underscoring the critical need for women's involvement in AI and advocating for diverse voices in training sets to ensure fairness.

Joy also highlights the transformative potential of AI to revolutionise women's roles in fields like marketing, urging them to leverage AI for enhanced skills and efficiency.

The discussion also tackles AI's challenges, such as data biases and fraud risks, and delves into ethical concerns about deepfakes and manipulated content. The speakers call for regulations and transparency in AI-generated content, especially in advertising and influencer marketing.

Takeaways

Women should embrace AI as a tool to enhance their skills and efficiency in various roles, including marketing.
Diverse voices and perspectives, including women, need to be included in AI training sets to avoid biases and ensure accurate representation.
AI has the potential to impact creative roles, such as marketing, by automating tasks like content creation and data visualisation.
Challenges and risks associated with AI include data biases, privacy concerns, and the potential for fraud and deep fakes.
Women should seek opportunities to learn about AI and its applications, whether through online courses or discussions with industry professionals. AI technology has the potential to create deepfakes and manipulate voices and images, raising concerns about trust and authenticity.
Regulations and transparency are necessary to address the ethical issues surrounding AI-generated content, particularly in advertising and influencer marketing.
Women's involvement in AI is crucial to ensure diverse perspectives and prevent bias in its development.
AI can be used creatively to enhance productivity, efficiency, and personal well-being.
Play-based wellness events and alternative networking opportunities can provide a more inclusive and enjoyable environment for professionals in the digital industry.

Links relevant to this episode:

00:02 - Joanna Shilton (Host)

Hello, and welcome to Women WithAI™, a podcast for women to share their stories, insights and expertise while also engaging us in conversations about the impact of AI on gender equality and representation. Today, I'm speaking with someone who's not only a commercial strategist, global ad tech startup pro, consultant, mentor and a public speaker. She also makes time to organise wellness retreats, hula hoop surf, teach yoga and roller skate. So I can't wait to hear more about all of that later in our conversation. But before we jump into discussing her passion for AI, let me tell you a little bit more about today's guest, Joy Dean.

00:36

Joy is an energetic go-getter with 18 years experience leading sales, data and partnership teams to success, and she's responsible for launching many leading ad technology companies during their reign across the digital ecosystem. As a commercially driven startup professional, her journey has been one of expansion and strategic connections. Joy is passionate about AI and its use in innovation horizontally across the tech and data landscape. Her ambition is to continually make an impact, leveraging her deep knowledge, dynamic personality and growth mindset, and she certainly does all of that. Joy Dean, welcome to Women WithAI.

01:13 - Joy Dean (Guest)

Hey, thanks for having me.

01:16 - Joanna Shilton (Host)

That's great, so it's great to have you here. I mean, so before we, yeah, well, to get us started, how did you end up in your current line of work?

01:24 - Joy Dean (Guest)

Oh, the long twisting and winding road of Joy ending up in AdTech. I actually went to university. I have three degrees. I went to university for graphic design and visual communication, and then I switched over to theatrical design and visual communication extension costume designing, and I've always been really interested in creating. I've always been really interested in innovation, whether that was digital, whether that was taking a, you know, something like a piece of fabric from a flat dimension from first dimension to third. I've always just been driven by that.

01:56

And in: 2006

02:26

I was already doing like costuming. I was running my own business like you're gonna get bored doing this. Like it's a junior role. I'm like, no, no, I love to learn. So it's a really. It's just another additional tool set that I'm bringing that I'll be able to kind of add to my wheelhouse, and surely, within about six months, I was already getting promoted because my skill set from working as an entrepreneur really was able to be applied to anything. I mean, that's one of the things about women that do different roles we are able to kind of transition our skill sets into other positions.

02:55

So I did, and eventually, the company started to globalise, and they needed someone that was not afraid to just run through the doors yelling and screaming about how great the tech was in another country. And so, yeah, they flew, they were flying me out to England, and I kind of launched the UK division, and eventually I was going back and forth, and eventually they were like, do you want to move there? And I was like, absolutely, let's go, let's do this. When they sent me to, they sent me to London, I think there were two other people in the company at the time and from there, I've literally been doing that same kind of move over and over again.

03:31

So, launching the US company, I brought it to the UK, and now I take companies that are outside of the UK, you know, whether it's in Europe or U or U? S and I help them kind of roll out their products or their strategy in the market. Yeah, and it's good; it's nice because it kind of leans into what I've always been doing, which is I've always been an entrepreneur. So it's kind of like being an entrepreneur, like a mini entrepreneur inside of a bigger bubble, right, because they hire me when there's no one else on the ground to basically launch and build, and that's what I've been doing my entire life. So, I just kind of transitioned it into advertising.

04:08 - Joanna Shilton (Host)

Brilliant, and tell me about your work at Tobii and how you use AI there, because, is it I'll get it right? It's eye-tracking, is that right?

04:15 - Joy Dean (Guest)

. Tobii has been around since: 2001

05:01

There you go, Alzheimer's or things like brain trauma.

05:04

So using eye tracking to actually understand cognitive response and behaviour and then using that to make analysis and inform decisions about someone's healthcare, I mean, it's amazing. So they eventually, in the last, as I said, six years, wanted to start providing this data to the digital ecosystem, and so that's where I came in. I came on board to launch a division called ATEX, which is an attention data exchange, and it's literally taking eye tracking, using eye tracking hardware and technology, to collect and measure data attention data and fixation and gaze data through the eye of an actual human on an actual page, on an actual ad, take that data as an output and then pass it on to a customer that uses it in their platforms to create their you know, a machine learning or AI-driven algorithms. So it's really the foundation of we're the enabler, of we're the enabler. We create the data sets that are given to others to then build on top of whatever kind of artificial intelligence that they deem fit or what's useful to their customers or useful in their long-term strategy.

06:15 - Joanna Shilton (Host)

Yeah, wow. And are you using AI to collect it as well, or not? You're the data; you're providing the data with the data.

06:23 - Joy Dean (Guest)

I mean, there is some kind of machine learning done in eye tracking in general.

06:27

So, eye tracking is when you have near-infrared light.

06:30

It's coming from a camera. The light reflects onto your eyes, and the reflection goes back to the camera. The computing system reads that, takes calculations and fixations, and then pulls that data to determine what you're looking at.

06:46

So there is a little bit of machine learning in this and the calculations of the information. But we are all about collecting the input. So this really high quality, passive, always on and locally collected data sets from an opted-in panellist, putting it, it comes into the platform. Our data system then gives, gives it an analysis, right, so it puts it into its correct buckets, right, if you will, in layman's terms, and then there's an output, and that output is attention, fixation, data against ad delivery, and then we pass that on to the customer and so yeah, so we're kind of the first step in the process of building an AI platform. We're the first step in the process of evolving something that you might already be doing by giving it more juicy, dynamic, sexy data points to then be able to build really robust and that kind of thing, like do you do equal amounts of like men and women?

07:45 - Joanna Shilton (Host)

and I don't know, does it matter eye colour, because I mean there's got to be differences with men and women like that's can. My worry this is what I'm trying to find out as I discover more and more about ai is that kind of. It's those data sets, it's making sure that that the, that it's equal data, that it's men and women, it's not being biased to one sex or the other. So, how do you overcome this?

08:05 - Joy Dean (Guest)

one of the inherent. There's always going to be inherent biases in some data collection, depending on how it's collected, because the whole idea is that you have to start for AI to work or any kind of machine learning. You start with a data set, and then you can, depending on who's inputting that data, who's collecting that data, they can transfer their biases onto the data centre. For sure For us, because we don't create, we are not creating the AI models; we are just collecting the raw data, so there's no bias in it. And because it's a panel that we're collecting it from, it actually allows us to see demographics, so we can see by age, gender, location, you know, whatever it is, we can see what your attention fixation is, so we can tell the difference between a man and a woman, a man and a woman of a different age, different age groups, different locations. So age, location and gender. It's really, really raw.

09:03

And I think it's a really good point to make is that when you're working with any sort of AI or machine learning systems, you need to make sure that the inputs are as high quality, as unbiased as possible because if you're making a prediction on top of a prediction, it leaves room for error and inaccuracy, which is why we always tell our customers and anyone that will listen this is me shouting off the rooftop here is to always make sure that the input is the highest quality, passively collected as as unbiased as possible, before you start using it to create your long-term ai, because garbage in garbage out, right like if you don't use the most high accuracy pieces, what you get on the other end is also going to reflect that yeah, no, thank you explaining that and that, because we are we're here on on Women WithAI and I want to you know, talking to all these amazing women like yourself, and what are your thoughts on the future of AI and how it might impact on women specifically, or do you think it will?

10:10 - Joanna Shilton (Host)

Well, as a woman, let me just start there.

10:14 - Joy Dean (Guest)

I mean, I definitely think there's good, there's a benefit and there's a. There's a benefit and there's the negative. There's that. That's with every technology. I think, inherently, data doesn't have. You know, in the, in the space that I'm in and data and tech, we don't have as many women on the forefront actually leading the strategy for go to market for AI, go to market for data or big learning or deep dive all of that stuff typically around data or coding or any of these kind of more technical areas tend to have a higher percentage of men working in the space, leading the space or being at a high decision-making level within the space, and that's something that we as women, need to. You know, as we spoke before, like lean into making a change, and you know, for me, yes, there should be, the ladder should be thrown down, of course, but also I am empowering myself and I challenge every other woman listening to this to empower themselves, to just immerse yourself in learning what it is and how it works and seeing if there's a space that you can slide into, because ultimately, the roles that are impacted with AI are typically more creative roles. Marketing will be heavily affected by it because now you can do. You know, use large language models to create imagery, to create text. A lot of the stuff that you would do as a marketing exec or as a marketing strategist which is like sales decks and creating collateral and creating branding all of that stuff can be done using AI. So those roles are typically female roles. That's gonna go, or it will at least have much. It will impact the amount of jobs that will be available in that sector, which means then you're fighting, you're all fighting for one spot, whereas if you can learn how these large language models work and how to use these inputs kind of call and response setups then it will make you more desirable as an employee. Right, so you already have the creative skill. This can then be an application of your creative skills. I think also the positive is you've got great opportunities for data visualisation, so making things that seem really difficult to understand to make them a lot easier to, because you can look at it. You can actually use AI to take data points and visualise them into something. Because you can look at it. You can actually use AI to take data points and visualise them into something that you can look at. Right, you can use it in so many ways. I mean even the things that we were doing previously, for you know an eye tracking, so healthcare and DMS systems and all of that. Ai is going to help a lot with that. So it's great.

12:53

But again, make sure, as a female, that we don't just assume that it's not for us, because it's all for us. It's ready, go pick it, go pluck the cherry, grab the apple. It's just about figuring out. You know how it works and that's ask questions, that's you know. Go online, that's sign up.

13:12

I mean, you asked me before about where to find this stuff. Like, I just started going to an. It's an app called Domestica. Shout out to Domestica and it's literally top level. You know online courses that are done in a more creative way. So if you're more creatively leaning, then something like going to an online course that's just reading a bunch of materials online so it gives you an easy way in. It's got lots of different, of different use cases and then if you find something you like, then you would go and, you know, maybe do some training. Or, if you're in a marketing role, already speaking to your employer and saying, hey, I'm really interested in learning more about how I can use AI in my role and start learning that way and seeing if you can get additional assistance and ask a friend yeah, call me DM. Me slide into my DMs, I'll help you yeah, no, I love that.

14:00 - Joanna Shilton (Host)

You're right it is.

14:01

It's about leaning into it and not being afraid, because I might I might have mentioned this before in the podcast I don't know there is, I think I did but there's a BBC article about women not using it.

14:09

I mean, this was last year and sorry for repeating it, but it I think it's really important because women, you know, as a woman, a woman you might think, oh no, I don't need that, I'm not going to do it, I'm going to you know that need to maybe sort of prove yourself. Sometimes it's like no, if the help is there, take it, use it and learn to use it, because actually, the more you use it, you'll get quicker. It will get quicker, it will do a better job of providing you with your social media posts or your marketing plan or your usually, you know, from a sort of marketing. That's what I use it for, kind of point of view. But listening to people, other people talk about using it and the more you ask it to create an image or something, it's about you knowing what you want it to give to you.

14:49 - Joy Dean (Guest)

So you're right, we can only use it if we do lean into it and 100% and also it makes it removes, because, remember, if this is a learning model, if this is a like, if you look at generative AI, it's about feeding in data sets and then it can learn from those data sets to then create new data sets in similar fashion, in the same kind of style. We need our voices in it, yeah, is in it. Yeah, so the more women that are using these kind of large language models or being involved in working in ai, it's getting our voices into those training sets as well. And that's like you know, we need to have women of all colours, we need to have women of all backgrounds, we need to have women of all coming from different regions, so we can also feed into these systems so they can benefit us. Yeah, right.

15:36

Or if you're you want to start your own business, like this is a way. If you're a mom and you're like a mother, right. Or you're a caregiver, let's say, or maybe you're a part-time grad student, these things, these models, can help cut time for the. You know that that's one of the best applications. It's like taking repetitive work and streamlining it, or cutting down time that it takes to get something done. Or, you know, giving you prompts to help you start something that is more difficult for you. Maybe you don't know where to start. These things can help you level up your efficiencies as one one in your business, but also in whether or not you want to launch a business or whether or not you need more time to do something else and just trying to find it, trying to think about it, less like this big scary beast and more like what could my day-to-day applications of something like this be? And then start there, because it makes it easier to learn from, because it's your day-to-day right, yeah accept.

16:28 - Joanna Shilton (Host)

The help is there. Use it to make yourself better. There was something that I saw this week or that's been out recently and made my blood boil. It was about going into chat GPT and asking it to do a kind of a job review, or kind of like. You know how someone done for the year, you know the kind of thing, and they put in all the details and they put in a man's name it was John and then they did the exact same and said can you do the kind of appraisal for the woman, for Jane? And oh my God, but it just it is just learning from what's been in there. So there's probably like a hundred billion thousand.

16:59

Okay, I exaggerate a little bit, but you know, good appraisals for a man and maybe there aren't so many for a woman at that same level. Just because there haven't been that many, it's not that women are worse or that men are better, it's just not there. But it just what's available. It just spat out something that was like oh yeah, john exceeded everything, he was great, did it? Oh j Jane, oh yeah, she met all the expectations. And it was like what? All you? But it's that unheralded yeah and it's the same.

17:22

You see it, that's why you don't put your name or your age, or you know, on job applications. Now you know, like, depending on where you're going, obviously, and whether it's for you know everything else, all the details need to be stripped back so that you can make because we, you know, we do all have that. I guess, that, well, unconscious bias, or it's just, it's very easy, isn't it?

17:42 - Joy Dean (Guest)

t's trained on data from like: 2000

17:56 - Joanna Shilton (Host)

like this.

17:57 - Joy Dean (Guest)

So the the information that it has to work from is the information that's out there. We live in a patrilinear society. We live there are there are. The expressions of, of representation of men and women are very different. So what it's learning from already isn't accurate, it isn't well, it isn't complete, and then there's fake news and then there's all kinds of you know, user-gener generated content that it's pulling from as well. So we have to think about those. I mean it leads right into the issues around it. The negative side of it. It's like privacy, you know, fraudulent information or just deep fake.

18:31

Or I was just reading an article someone shared in another group that I'm on and it was about the one of the senior leads at WPP having a deep fake attack on him where someone had used AI to clone his voice and then went to try to yeah, to to open accounts with clients, to get them to invest money. You know, or other people that have been. You know, using the voice activation or the clone voices for the bank like getting people losing. You know, using the voice activation or the clone voices for the bank like getting people losing, you know, millions of dollars. And I literally just it was maybe a month ago, I saw the release of like an updated voice cloning.

19:08

You know technology and I was like, why would you ever? I don't, why would my voice, why would I need to clone my voice? And people were like, oh, but you can leave messages from beyond the grave, or or you can use it to make that. And I'm like, for me, if you, if I clone my voice and also I watch a lot of sci-fi. I watch a lot of sci-fi with robots, like pretending to be you and like taking over your life and I just think it opens up opportunities for a lot more fraud than it's necessary, than the benefit yeah would would bring. Yeah, just because you can, doesn't?

19:39 - Joanna Shilton (Host)

mean you should do that exactly. We've created all these things have been created that could do this, it could do that, but should it should do, we need it. There was. There was a drama bbc drama, maybe it's like last year, the year before, I can't remember the name of it, we'll find it but it was about people changing the cctv cameras. So you know, crime's being committed, someone's sitting on a bus or someone's in the street, but they were actually changing it. But the thing is it seems far fetched, but if they're doing it you can guarantee it's being done or someone's going to work out it. But it was. They were changing the voice. They could change the person's whether they'd gone on the bus, got off the bus, done.

20:14

I'll try and find what it was called. But it's just, it's scary that it's doing that, and then you don't know what to believe. And then the person's on TV and they're recording a live what they think is a live, you know interview and actually it's not. There's the delay. They've changed it. They've changed what they've said.

20:29 - Joy Dean (Guest)

It's terrifying. So I mean they do this all the time in Mission Impossible, changing the fake face. I mean, it's the same, you know, I mean it's interesting that I mean, and it's fun, it's funny, but it's also really it's. It's interesting because technology, with all of its benefits, it always has these kind of byproducts of negativity, and there's always scammers.

20:47 - Joanna Shilton (Host)

In anything that you do, there's always well, yeah, I mean a signature. Really, what does that prove?

20:52 - Joy Dean (Guest)

right. So I think I do think that there has to be. We're still in the this is like the rise, right. So we are in a time where we still need to put protocols in place. We need to make sure that all the data collection so if you think about us with eye tracking, where we look because we are able to track eye color, eye gaze, eye shade, if you're wearing glasses, if you're wearing a hat like to get really, really accurate information. Now, if you don't have privacy and protocols in place to protect that information like we don't share any personal identifiable information, we don't share anything that's about the user, just about an aggregated group number dah, dah, dah, dah dah.

21:27

But if you, if, if you were to have your hands on data sets that could alter which generative AI does it can alter the appearance of something or create I mean, you can create fake people and they actually sound like real people and look like real people. There's a woman, there's a agency that does it with them. They do a Spanish agency and they do, oh my God, influencer marketing, where they created a fake person. Yeah, she's a model. So there's a huge rise of these AI models. So you don't have to pay a human being.

22:02

You have a fake person that looks like a real person, like it's getting so good, and then it can, you know, sound like a particular person or they've made up a new, a new voice, and put that out and it can be part of your marketing strategy. There's companies using it for campaigns. There's, you know, brands that are testing it out for advertising. So there needs to be obviously some regulation. So it says. I personally think it should say if you're, if you are selling something or doing an ad campaign or whatever, and it's representing a person, but it's not a human person, it should be defined or should be marked in that way, you know, like this is an ad using a ai generated model, something like that. So we can start to maintain control of what we're seeing and what we're using and not just be completely in the dark because then you've made your own decision as well, whether to trust it or believe it or go with it.

22:59 - Joanna Shilton (Host)

You know, and we do. You know you trust the famous person that's maybe saying this, but you might get someone that's promoting the vegan sausages and then you find out that actually they eat meat and you're like what? But you've lied to me, you know, and it's like the same. You might be thinking, oh, I really like that woman or that man or whoever I'm following, and you're like oh, hang on, they're not real. I mean, there's someone behind it, but it's not, it's a team of people.

23:20

But then that's the same with that. You know all famous people, you see them. You know we've seen all the pictures from the Met Gala recently and. But you know if you see behind the scenes, they haven't just stepped out of the car looking like that. There's a team of people behind them doing the hair, doing the makeup, blowing you the dress in the right way or the tuxedo Holding the cake. Yeah.

23:39 - Joy Dean (Guest)

I mean, by the time the images reach us, they've been airbrushed. I mean, can you think about it Like? Ai is just another level of what we've already had.

23:46 - Joanna Shilton (Host)

It's just another tool.

23:47 - Joy Dean (Guest)

It's like Photoshop is what I studied in school. Right, and you can warp the body, change the body, change the image, edit the image, and this is is an extension or a leveling up and an enhancement, and and of that it's it is creating. It's kind of what we've always done. It's kind of a human thing to do. It's like we're always trying to be better and make our tech better. So let's figure out how all of us can use it for the better, but also to keep it clean and know what we're, know what we're, what we're using and when it's being used, and that's the important bit about getting more women involved, or using it and inputting into it.

24:25 - Joanna Shilton (Host)

So it it is.

24:26 - Joy Dean (Guest)

It is more representative.

24:28 - Joanna Shilton (Host)

Yeah, it's okay, that's all the. That's all the bad, scary stuff. What excites you about AI, like how, how do you use it or can you? Yeah, do you use it in your? What are your passion projects and?

24:40 - Joy Dean (Guest)

I mean this is the same things that I mean the same thing. That is can be scary about it are the things that also excite me about it. Like I love, I'm an. I went to school to be an artist, I have an MFA, so I love creating things. So I love being able to sit down like, if I take a photo of something I don't know, I've got like hula hoops right here next to me, right, and I take some pictures of my, my studio. So if I'm like hula hooping and I take a photo of me hula hooping and then I import that into a platform, and I start adding on detail to it, I can put it into like whatever world. I can turn it into steampunk image, I can make something from my own work and then I can embellish. Now, the scary part of that is always like someone else taking your imagery and then turning it into something and you don't get credit or things like this. But I love the idea that you can create alternate worlds. I love the fact that you can use it to have efficiencies around, like I use it for sales decks and collateral brand pitches.

25:39

I'm trying to build a festival in Portugal at the moment and I use it for helping me understand how to do an outline for fundraising. That's just like stuff, like largely just using something basic like a chat, you know, like a chat GPD, like how, giving me an outline of how I can actually start to build my fundraising products and getting prompts, putting in prompts and getting prompts back about things that maybe I didn't think about adding to because I've never done it before. So that's quite good. Or time-saving, like all the repetitive data entry that I have to do from any kind of either a conversation that I've had or something that I've sold, and instead of I got my notes here, instead of doing that, I can record it and it can just set it up for me, pull it out by category for me. I don't have to do that.

26:25

Manual work, manual work. So it saves me time. I think I most like, mostly I use it in more of a marketing context and I use it more in a data organizational construct. I use it in a visualization construct because, again, I work with a lot of data points at Tobii, so when I get outputs I like to see it and we don't get. These are the things that we give to clients, but it's things that I have for myself. So when I'm looking at data, it can be it's easier for me to understand it, but also it gives me ideas because I see it in a different way. So, yeah, I I'm much more of a use it to learn and then also to create fun things for kind of my own passion projects. So tell.

27:12 - Joanna Shilton (Host)

I'm fascinated to hear more about the festival. But how has that come about? Because I mentioned earlier, right at the beginning, because you do wellness retreats, which I think at the moment are just for women. But that's just for now, isn't it? I don't know.

27:23 - Joy Dean (Guest)

That's for now. Next, we got to get the men in there, but that'll be the festival. Yeah, tell me that'll be the festival. Yeah, tell me, tell me about that. I want to hear everything. I mean I, like I said, I've always been a creative and I always love making stuff. So in part of building and creating as an entrepreneur, I'm always looking for other avenues to to do that in. I get exhausted by the idea of the only time that I can do a client meeting is if we're doing it over lunch or if we have to go have a bunch of drinks and because I'm getting older, so you know, drinking like takes two days for me to recover from so you know what I mean.