Artwork for podcast Women WithAI
E8 - Balancing AI's Influence in the Media Industry with Fern Potter
Episode 822nd May 2024 • Women WithAI • WithAI FM
00:00:00 00:29:03

Share Episode


Fern Potter, Senior Vice President of Strategy and Partnerships at Multilocal Media, discusses the use of AI in the media industry. She highlights the importance of data quality and ethics in AI models and the need for human oversight, she emphasises the need for AI to be used responsibly and for the data bias to be addressed, she discusses the impact of AI on journalism and content production and the challenges of maintaining privacy in the AI era. Potter concludes by expressing her excitement about the potential of AI and the need for a collaborative approach to ensure its positive impact.


  • Data quality and ethics are crucial in AI models to ensure accurate and unbiased results.
  • Human oversight is necessary to mitigate risks and ensure responsible AI use.
  • AI has the potential to revolutionise journalism and content production, but careful consideration is needed to maintain trust and authenticity.
  • Privacy is a significant concern in the AI era, and regulations are necessary to protect individuals and businesses.
  • Collaboration and a holistic approach are essential to harness the full potential of AI and address its challenges.


00:00 - Joanna Shilton (Host)

Hello and welcome to Women with AI. Today. I'm thrilled to have an expert in developing productive partnerships across the media ecosystem on the podcast, With extensive experience in strategy and partnerships. I'm looking forward to her sharing her perspectives on using AI to respond to the changing needs and requirements in this fast-moving media industry. But before we jump into our conversation, let me tell you a little bit about her.


Fern Potter is Senior Vice President of Strategy and Partnerships at Multilocal Media, a pioneering curation business. Fern has spent 16 years in media, the majority of which in programmatic. She has a firm grasp of technology, data and its application to benefit advertisers and bridge the gap between consumers and their experience with different brands. Having previously worked in media agencies leading the global digital strategy for notable brands such as JLR, Diageo and VWG, understanding advertising and buyer needs and how ad tech can solve for them is her speciality. Verne is motivated by advancements in technology and data and in her current role, leads strategic partnerships globally for Multilocal, adding value across the programmatic ecosystem in bounds the programmatic ecosystem in bounds. When she's not meticulously picking through the latest industry advancements, she spends her time in the countryside with her sidekick Blue, her gorgeous yellow Labrador, and is apparently an avid F1 fan. So, Fern Potter, welcome to Women with AI.


Thanks, Jo I love the introduction much appreciated, great to have you here Before we get. Well, to get us started, what exactly do you do for Multilocal?

01:28 - Fern Potter (Guest)

Yeah, absolutely so. Multilocal is a curation specialist. The business has been around five years and essentially we work with various partners across the programmatic ecosystem to add value to their media campaigns through programmatic, to add value to their media campaigns through programmatic. So, whether that be DSPs, ssps, agencies and clients, we build partnerships, which I work across globally to make sure that campaigns are performing and the optimisation is working. So we have an extensive network globally and a great team across the world that are dedicated to providing that kind of partnerships and customer service. In the new world of curation it's a relatively nascent industry and so it's really exciting to be on the forefront of that as it develops and as there's a lot more noise this year around how curation can help solve for various different kind of programmatic inefficiencies that have existed before.

02:28 - Joanna Shilton (Host)

so and you're, I'd imagine it you would use a lot of complex, uh, complicated, algorithms. So how is ai sort of impacting on that, or have you always been using ai or how do you use it?

02:41 - Fern Potter (Guest)

yeah, it's really interesting, jay, because actually I was talking to a few kind of colleagues and peers when we were having a conversation about appearing on the podcast and I think the interpretation of AI is actually quite vast. But that said, you know, look, we've been in the ad tech industry itself, been using the likes of algorithms and machine learning for a long time, and I think AI is a generative step in a direction to really bring about more sorry, to kind of ingest more data and more complex data from this kind of 2D environment to a real 3D environment to bring about decisions and change. So in our business we've been working with, like what I would say using the iceberg analogy, the kind of underneath all that data processing, data scrutiny, analytics, you know, data interpretation, data completeness, all that framework and foundation that really is required to have rigor in what the AI models are then built upon, and I think this makes for better decision making and more trust and confidence in the outputs that AI models are delivering into business. So we've seen real change. It's gone from essentially perhaps you know most users are categorization and optimization tool into real world kind of examples and almost around this kind of executive assistant where you can build creative decisioning and find opportunities in data that you might not have once been able to do when it was more of a kind of standard algorithms and machine learning protocols, so it's been fascinating to work in that development.


We're talking about AI in our industry as something that has been ongoing for a while but has been a relatively new hype curve. If you want using the Gartner hype curve as an analogy, last year it really exploded with ChatGPT OpenAI Baird from Google, and I think that has created a lot of opportunity in our business to help with productivity drive results. But I think top of mind for people is always understanding the data set that this decisioning is based upon, and that's really important, because how do you make sure that you've got the right data?

05:04 - Joanna Shilton (Host)

Are you seeing any bias problems or anything like that, or where does it come from?

05:10 - Fern Potter (Guest)

yeah, there's. I've had a couple of experiences with that, joe, I think. Um, what from a? From a business perspective? Now we work on a lot of data.


Like I said, the company's been around for five years and we're using ai to help categorize our audiences, but that's based on five years of having a team in place to scrutinize that data, and I'd say about 60 percent of the data that we're working upon needed and required, you know, an assessment and a quality assurance to make sure that it was then fit for purpose, based on what we needed to do as a business, and so the team have been learning as we go on how to make that data more solid, more consistent and reliable, so that the AI that we build into our application to make things quicker, better and stronger is based on something that has a confidence level that we can be assured, because one of the biggest issues in media is wastage.


Now, if we're making decisions on bad data, you're kind of exacerbating that problem for your clients and your businesses. So we've been working hard and there's still a human element. There's a lot of QA that goes into um the outputs that we get um, and it's a you do, you absolutely do um, I think you know we'd be.


Let's learn from experience. We've we've spent a lot of time relying on machines, whether it's from a verification perspective, fraud detection, um, the categorization of youtube around 10 years ago, where there was there's a misplacement of how machine learning was, was categorizing that, that inventory, and now, um, we don't want to replicate the same dysfunction. So I think there's a real opportunity to making sure that, as I said, there is rigor and there's confidence in the decisioning and then there's human quality assurance so that, essentially, brands, at the end of the day, aren't aren't wasting media dollar, but also you're providing a brand safe and suitable environment for advertising and and that's crucial to that brand experience. And the other point in experience, I worked in a business several years ago, working for a creative agency, and when I talked about completeness of the data, I think there's definite bias in what we have to work with today and perhaps what we will work with in the future, and there's a real my experience of that is I worked for a business called FCP6.


My experience of that is I worked for a business called FCB6 and within that, they created a utility called Black and Abroad, which was for domestic travel for the black community within the US. There was no record from government or on systems from national parks et cetera, of cultural significant locations specifically to black people, and I think that was something that you then see is like an inherent bias within the data that we're collecting. So it's not only the individuals that are creating the data, but actually the register of how this information exists already.


Yeah, do you already, yeah, and so that Do you know. Yeah, and you've got to have that kind of keen eye to kind of. I think that was a blatant miss based on Because the business we were working with, that's what we were creating the utility for. But you could unconsciously miss some of that data bias by just working with what you've had, miss some of that data bias by just working with what you've had and um. So making sure and you know there's a huge opportunity there which feels quite beautiful, and bringing like a cto and an ethics together, I think that's that's going to be fascinating over the coming.


I always say years, but probably looking at months, to be honest, aren't we to like, I really understand the quick, yeah, the rapid growth of this business and um, but to really understand, like, how those, how ai, can exist in a positive way? I think it's those two kind of brains within a business that come together and have here's the ethical leadership, what that means for data biased, and here's the technical leadership and um. I think that I find really fascinating. Actually, it's quite a.

09:27 - Joanna Shilton (Host)

It's a different shift. It's like you need the humans to spot where the data is missing, so it might not just be like that the data is incorrect. It's actually there isn't that data because it hasn't been collected or it hasn't been saved, and that's what we're finding out, so I guess that's the kind of it could be a negative, but it's also a positive because actually it's highlighting, as long as you've got the human element, checking what the data is and where it is? I mean, yeah, I mean, what excites you about AI?

09:57 - Fern Potter (Guest)

Oh, you know what I mean. Like I said, I get really excited over every technology, but I do feel that there's something here that it's a generational shift. I was reading something on Forbes about has been the transition generation, because we're kind of seeing AI as how it fits into our culture, so not only as a business, not only in ad technology, but as a culture of how humans interact and work. And I think that transitional shift is interesting because, say, in 10 years' time, it's just going to be intrinsic in everybody's way of working. And so we're on the cusp of setting that up really and we're at that kind of leading edge of understanding what's right, what's wrong, what works, what doesn't, and mitigating risk, as much as we are spotting opportunity. And you know, I think about my role within a business. As much as strategy is about finding opportunities and partnering with people to realise those, it's also about mitigating the risk of a business and one of the things I'm quite excited about when it comes to AI which does sound quite nerdy, but it's about continuity. So when businesses need to be smart and they need to move fast, intellectual property and experience often sits with the individual. I think AI, from a gener perspective really would help us centralise the intellectual property that sits within people into a business infrastructure.


What we suffer from in business, and especially in ag tech, is shorter tenures, shorter tenures of individuals. You know, a C-suite might have a tenure of around four years, but actually people at the cutting edge of what's happening on a day-to-day basis with clients and within business maybe one to two years. You have disparate systems, whether it's project management, whether it's your LinkedIn communications, whether it's a CRM. So, or you know, someone saves something on a desktop instead of a SharePoint, you know, god forbid. So you know, being able to centralize all of that intellectual property into something that can provide, you know, de-risk a business in terms of having the continuity um outside of people, I think is a real, is a real smart move for business in um being able to keep propelling forwards.


And you know, I imagine a world where this there's a, you know, there's a brain that sits in the center that can, you know, that humans act upon but that can actually tap into historical positive experiences, but also negative ones. I mean, unfortunately, as humans we carry ego, so we are a bit biased to what's good and talking about what's good and what we've learned in terms of, you know, positive results, positive experiences. So there's a bit of a record loss on what didn't work and what didn't go as expected, and that's a really valuable experience as well. So if there's something you can kind of dehumanize and create a centralized brain, I think for continuity, that's really smart business management, especially, as I said, in a world where tenure is shorter than it used to be. So that kind of excites me.


One thing that I'm not, you know I'm not an expert in at all, but I'm very interested in how AI is going to kind of influence journalism and content production. You know that goes back to ethics and trust as well, so I'm kind of watching that evolve. You know, do we need to be highlighting what has been written by AI versus what's been written by a journalist, and what does that mean for news and communication on a global level?

13:42 - Joanna Shilton (Host)

and I think that's um, that's quite, I agree.


That bothers me as well, because you, you know, you hear about a generative ai being um, like hallucinating sort of making stuff up and and taking it, and I think, I think, yeah, I think everything should be labeled, whether it's purely ai, whether it's sort of a combination of human and ai, but because humans get it wrong, just like AI gets it wrong, like we had an instance, uh, just this week, where an article was written about a project that someone, a journalist, just thought, oh great, I'll just write about that.


And they did it. But they'd taken all the wrong details and it was the wrong information that got sent out. And we're all panicking where's this come from? What's happened? And, um, you know, by speaking to him it was like, oh well, actually, yeah, I just took this, I thought it would be useful, did that, but that that could easily happen with ai. So you need to, you still need to have the human element and, as you say, if you've got a sort of central database always gathering all that data and it's freeing up time, then that's going to give us, like the human, more time to be doing products that ai or the machines aren't capable of.


You know, because we've got that expertise still you know, until ai takes over and becomes more expert, but I don't think it will. I think there's different levels, but yeah, so, yeah, I think you're right, but you know, where machines are unable to match human performance.

14:58 - Fern Potter (Guest)

We need to make sure that we're just not letting them get carried away it's true, it's funny, I don't I've used um, so I, you know, I transcribe my meeting notes and then I can use like a chat, gpt or something to then, you know, turn that into an action plan with key actions, like that's really useful. However, getting that done and and having communications and partnerships with individuals and businesses isn't going to be replaced with ai, because there's a human connection and a kind of a coming together of and collaboration on ideas and what you need to solve for. So ai might be able to provide solutions, but you've got to tell it what your problem is and I think I I think that we're still going through that transitionary phase. To your point, it's not human decisioning still needs to happen, human relationships still need to happen and you know, the data that it's working upon at the end of the day was inputted by somebody, so it's like as much as it might change and start to self-learn. I think you know that comes back to the bias and ethics. Like what are we talking about? And I was reading the us government. They've actually said for each of the government departments they need kind of an ethical ai lead um.


One of the hot topics at the moment with the, with the cookie deprecation is about privacy and you know that's huge for AI. Like, how do you protect your privacy as an individual? How do you protect your privacy as a business, as an entity, as a brand? And I know that the 4As did a quite extensive paper on this recently just to be able to kind of write into contractual obligations. Privacy around AI models because they are learning what they're learning from and what we're inputting into them needs to be kind of regulated. So that'll be interesting to see how privacy plays out in the world of AI, because you know, like with any new adoption of technology, there's kind of resistance, there's's push pull, there's cynicism.


We talked before about that garden, that height curve. It's like, yeah, it's really exciting. But now I'm in a bit of a maybe the trough of disillusionment where you do follow. No, there's a bit of cynicism. There's there's resistance to cultural change, but also defining what is, what is ai and what isn't um, articles around grifting come up quite frequently whereby you know people might say they've got something that they haven't and how do you unpick that? And um, which is, which is not good, not good for the technology or the industry. But there's, uh, as I said, those frameworks, that ethics, really digging into businesses and understanding what they're building those models upon and how their models work. I think it's really advantageous for anybody that's working with a business, with AI or an.

17:52 - Joanna Shilton (Host)

AI-led business, because the world of cookies completely changed, didn't it? And I suppose it's trying to make it better for the customers, you know I mean. So how is AI sort of affecting how people see advertisements and how they get channeled?

18:09 - Fern Potter (Guest)

Yeah, perhaps I mean I know, definitely from a creativity perspective, that you know we're having an influencer perspective. There's a lot more technology that can help build creative quicker and do decisioning on the fly. Before when, like, dynamic creative optimization came into play with programmatic, whereby you know there still needed to be an approval system for something that was automated, you know, to make sure that brands are happy with that content. So how that looks, I think there's still an approval issue, that that AI might be able to do it quicker, faster, better, but not if there still needs to be a kind approval, um, um and guidelines and and um hurdles to to move through um.


But I talked to my mum about it to be honest, because also, you can get really caught up in this, you know, in the industry and in the spin and forget that actually as a as people, as a general public, like what do? Like what do they expect out of that technology? How do they feel about it? Which was the same with privacy, right, when all the cookie pop-ups came up, you're like what does that even mean when it comes to consent? So, yeah, my mum was saying I think to your point before that you mentioned, jo, it's like you know, I want to know if AI has written an article. If the content I'm reading is from from a journalist, maybe she's more likely to follow journalists than she is to follow titles. When it comes to the, you know, her consumption of news, um, and so that might be kind of a fundamental shift which could be positive um. But yeah, you know there's a there's a lot to kind of work through.

19:53 - Joanna Shilton (Host)

I think, as we said, kind of culturally and and as an industry speaking to a bot, because quite often you go on something, oh, chat, and you do it and then you realize, oh no, this isn't an actual person, this is just giving you some standard answers. But the thing is, you more you confuse it, you know that it will eventually go. Would you like to chat to a real person?


you're like yes, yes, I'm not going to give you any of the how do you do use ai in your I was gonna say in your personal life, but I guess you know if you've got a siri or an alexa you do.

20:21 - Fern Potter (Guest)

But yeah, I'm not. Do you know what? I'm not too good on the, the voice activated um technology. I always have this fear that everybody listens to everything. But you know there's definitely. It's like a personal assistant for note taking action plans. I mean it can speed all that up and it does the heavy lifting, which I think is really beneficial, and you know we do in business. Look at it, for you know suggestions. So when we're curating inventory, we can layer the AI technology within it to kind of find suggestions based on what we're already running um. And that categorization is powerful but, as I said, only if the data is correct that you've put in. And it still has to be quality assured by by a team of actual people. So, as much as there's ai, there's also eyes.

21:07 - Joanna Shilton (Host)

Well, yeah, and it's getting that data, isn't it? And that's where like coming back to like, as you said, like data sorry, bias and ethics, and it's like if there are whole chunks of data missing and that's my worry and that's you know sort of you know. This is women with AI, and it's with amplifying the voices and perspectives of women working in the field. But also having that kind of underlying thought is it going to affect women more than men, or is it about race? Is it not about gender?


And I think you're right, it's just making sure that that data is there, but I think that's why you need men and women and all races, all nationalities kind of checking it. You can't just let it run off. Yeah, otherwise you're right, we're not riding the hype car yeah, absolutely. We're in the trough of disappointment.

21:51 - Fern Potter (Guest)

the hype curve are we in the trough of disappointment? Yeah, but we can change it. It's so true if you, we can, and I think that's the main thing is it's like, if you, if there's a notable observation that this is a risk that needs mitigating, to make I, I, a, I make AI work for everybody and make it the powerful tool. It could be that, you know, these observations need to happen now and they need to be solved for now, and there needs to be people that are actively ensuring that data bias and ethical use of data ensures that AI becomes, you know, a global platform for people. Like we said, it's a cultural shift as well as it being a technology shift, and so if we've got people looking at that now, I think that poses puts us in a good position for the future.

22:47 - Joanna Shilton (Host)

It feels like this is, yeah, an exciting time, I think I hope this is, yeah, an exciting time.

22:56 - Fern Potter (Guest)

I think I hope yeah it is. It's like you know, don't say you're on that. You're on that cusp, you're on that bleeding edge of change and you know that we're probably going to feel that for a while and as much that got in a hype curve will probably happen on a daily basis with some something you've read or an application you do that you've seen of ai that you disagree with. I mean, I was talking in somebody raised.


There was an AI beauty pageant that one of the gaming companies had launched, which was for women, and it's just disgusting, to be honest, to think of like AI used in that way and there was a prize reward, and you're like, seriously, have we not got past this, I mean, and so I think, pointing out bad actors um, that's a real blatant one, um, when it comes to to bias and ethics, but also, um, again, from a for a business, even even good actors, just really having your eyes open and scrutinizing that application and what, what the data is built on, because you know, a lot of the data isn't owned, it's public information and so, um, how, how you know anybody that's applying ai? On top of that, how are they making sure that day is complete yeah, it's calling it out, isn't it?

24:09 - Joanna Shilton (Host)

and, like you said, because I worry about that, you know I've got two young nieces and it's that what, that pressure that is going to be on them, what all young children, whether you're male or female. And did you see the Dove advert? Recently They've done a kind of new advert saying you know what is beauty? And it's showing all this AI generated beauty content and how you know soul destroying and devastating it can be if you don't look like that. But actually you need to show real women. I'm just using women now because that's what the dub advert was focusing on, you know, for their moisturizer and what have you. And, and it's just really refreshing I mean, they were actually then using ai generated images in the advert rather than real women so interesting, yeah, not quite, yeah, reality kind of that fake ai, and I think the thing is the more you look at it, the more you spot it.


But a young child doesn't have that experience, hasn't looked as as many images as you or I have, or anyone over a certain age that you're just not used to it. So you probably do look at them, think, oh goodness, is that how my skin's supposed to look. Is that how my eyes are supposed to be? You know, sparkling, sparkling. So that's the scary bit.

25:17 - Fern Potter (Guest)

Yeah, it is scary, it's a concern, isn't it? You think actually we've been like touching up photos and Photoshopping images for quite a while. I've done that myself. I was like wondering can I have a filter on? But there's that evolution, like you said. You've got've got, I think, in your mind. From a generational perspective you know that's happened, um, but if fast forward 10 years where everything's kind of self-perpetuated, ai, how do you have that critical analysis of going actually?


people don't look like that I know you could probably just go outside and you'd have some mates and stuff. But there's something in it that is it plays on the human, psychological kind of nature of of um of people and, and you know, we see it now with young women and, as you said, young, young boys as well um having image issues because of what they see through social content. Um is that is? Does that become more of a mainstream problem outside of the walled gardens? Who knows um? But there's definitely there's a responsibility there isn't there for for all of us um to to make sure, yeah, there's, there's a reality well, yeah, no, I think.

26:28 - Joanna Shilton (Host)

yeah, we need it to help us not just take over, and we shouldn't be aspiring to it, it should be aspiring to us.

26:37 - Fern Potter (Guest)

Yeah, that's quite right. Actually, that's a good way, a good perspective of it. You know it should for the greater good. There's so many awesome applications, you know, outside of ad tech. If you look at big enterprise businesses like the NHS or something, you look at surgery, you look at agriculture, like the way AI is shifting those businesses is fascinating, I think. I think, you know ad tech is probably a small part of how we can deliver real benefits, but there's always going to be those kind of hurdles along the way of it being used, um, inappropriately applicated, I guess that's been fascinating, like have you, do you do any sort of research yourself, or have you got any recommendations for places for people to sort of learn about, about what you do or about using ai?


yeah, I mean we, um, we use it as business, we're using it as an application into our main core business. It's not something we're kind of integrating it as we go. As an individual, I do a lot of research and I'd say, you know, actually the best people and resource that I have is the ad tech community that we just talk to each other. You can have those honest conversations about what works and what doesn't and, as I said, just hearing conversations this morning about what is the definition of AI is really fascinating. So, in terms of recommended reading, I mean Microsoft Copilot is really fascinating and, looking at enterprise level AI companies, I think they're on the cusp. It's also good to read some of the kind of edge cases and a bit more of the conspiracy theories around AI, get a holistic view.


So you know LinkedIn's always a valuable asset for trying to find that information.

28:32 - Joanna Shilton (Host)

And people can find you on linkedin, I'm guessing yeah, they can.

28:37 - Fern Potter (Guest)

Yeah, thank you so much for coming on women with ai oh, thank you so much, joe, it's lovely.

28:44 - Joanna Shilton (Host)

Thanks for having me.

28:46 - Fern Potter (Guest)

Thank you, you too.





More from YouTube

More Episodes
8. E8 - Balancing AI's Influence in the Media Industry with Fern Potter
7. E7 - Women Leading AI Innovation with Catherine Breslin
6. E6 - The Impact of AI on Social Media: Insights from Social Media Consultant Louise Brogan
5. E5 - The Symbiosis of Technology and Human Touch in Storytelling with Natalia Talkowska
4. E4 - Harnessing AI: the evolving landscape of tech, creativity and female leadership with Julia Linehan
3. E3 - Exploring the Intersection of AI Voice Technology and Gender Dynamics in the Workplace with Izabela Russell
2. E2 - Empowering Women in Tech: Harnessing AI for Gender Equity and the Future of Work with Sophia Matveeva
1. E1 - Introducing Your Host, Joanna (Jo) Shilton
trailer Welcome to Women WithAI