To AI or not to AI: a humanitarian comms conversation

Please note: this transcript is generated using AI

Welcome to this episode of Fresh Humanitarian Perspectives. My name is Wando and I’m delighted to host episode which is a light-hearted yet serious conversation about how we are communicating as humanitarians and comms professionals in this AI era.

Fittingly, I’m joined by two highly experienced professionals, David Girling and Debbie Adesina, who conducted the interesting research. We’re really reacting to the news article you wrote and the research that you and David conducted, on images and direct mail campaigns from charities. In this episode we’ll be looking at how and if AI images might be a tool organisations can consider now or in the future. And we’ve called this episode To AI or Not to AI: A comms conversation.

David
Okay, first, hi everybody. My name is David Girling. I’m an associate professor of Media and Global Development in the School of Global Development at the University of East Anglia. I consider myself a pracademic. I’ve been a Practitioner for many, many years, working in communications, marketing, pr, media, and I’ve been teaching in the School of Global Development now for about 13 years. And so that’s sort of a mixture of theory and practice.

And I still produce newsletters, I still help out with social media, I still help with films, etc, and I’m really, really interested in ethical representation and how unethical storytelling, how images are used, how storytelling is used, how film is used, etc. So that’s a bit about me. Debby, do you want to introduce yourself?

Debby
Oh yeah, hi. Hi everyone. My name is Debby. Debby Deborah Adesina. But yeah, I go by Debbie Wonders. And Okay, so now I’m a little bit stuck where my, my journey into ethical storytelling or you know, representations of global poverty actually was solidified or I would say maybe has its origin in my degree in Media and International Development, Masters in Media and International Development at the University of East Anglia where I got to meet David and I was introduced to media theories, media in practice.

And I also got the amazing opportunity to join David on the charity advertising project. And then that led up to joining Tear Fund, the international development organisation and just seeing the practical outworking of these tensions about what we see say about policies and guidelines and then practices, you know.

So that was an exciting intervention so to speak. An exciting experience of seeing ethical storytelling in practice. And I’m also looking forward to how strands of this will also unravel in my doctoral research on Nigerian celebrity humanities at the University of Liverpool.

And this is because I think oftentimes we binarize ethical storytelling as white presentation of black or brown bodies, whereas there’s a whole critique of privilege and poverty that applies to every single one of us. Black, brown, white, blue. Yeah. So looking forward to talking, tugging on those threads.

Nwando
Your PhD research sounds very exciting. Sorry, can you say what the topic is again?

Debby
So it’s. I will be looking at celebrity humanitarianism, but from the perspective of the global South. So how Nigerian celebrities do humanitarian work, how they you know, the performances of it, the politics, the processes, the impact of it and you know, just sort of contributing to the de. Westernisation of celebrity studies as it were.

Nwando
Very interesting. Yeah. A part of your research is mentioned in the recommendations that perhaps that could be, you know, something that people can consider is telling the story, flipping it and telling it from the side of celebrities.Your research will be really interesting. Before we even go into research, because I’m really interested in how you and David went about the entire process of what you did and what your thinking is now, especially with regards to the topic of AI localisation.

I wanted to ask this question around, comms today, David, especially with you, where you’ve been teaching for 13 years, which is by now you’ll be a teenager if you were a baby when you started. But your perspective on this, I think would be really interesting to hear is how has a visual. From your experience or from what you’ve seen, how has visual storytelling really fared with evolution and how things have changed with tech and all those things. How would you speak about comms today? Especially looking at our side, Humanitarian and then obviously private sector.

David

So before I actually started in the School of Global Development, I was in charge of digital marketing, I suppose you could call it, across the university. And I was very, very interested in social media, how, how social media, could be used, to support students in that particular context. And we had a great intern that did some absolutely wonderful things with the university, changing the way really that we told stories about the university. And so when I moved over to the academic side, I started a blog called Social Media for Development, which, which still exists, but I haven’t blogged for about two years.

In fact, I think Deborah helped me with that last blog post. So social media really has changed and I don’t think it’s the panacea of communications. In fact, sometimes I think that there’s too much emphasis on social media. There’s a whole range of communications that charities and NGOs still use, from anything from, an annual report to direct mail, to websites, posters, TV adverts, billboards.

There’s lots and lots of different things. When I was in Nairobi last week, was it. I can’t remember now, last week, I think, and I was in the, the Sarat Mall, and I came out of, the Carrefour supermarket and there was SOS children, and they had a banner and they were fundraising. And I found it really, really interesting the kind of imagery they used. And I was, oh, have you got any leaflets I can take away? And they said, oh, no, we don’t we don’t have leaflets. And so I think it’s evolved by using social media. It will evolve even further with AI, and there are some criticisms of AI in fundraising and development communications at the moment.

And it will just continue to evolve. But I think people shouldn’t forget the traditional forms of media, of pr, the importance of pr, the importance of prints. People actually still like print. They like to read print. But the images. There’s this constant debate on, you know, should you use negative, positive, post humanitarian, so, or no image at all. And I think that that debate has really, really, really changed since COVID since Black Lives Matter movement, since the decolonization movement.

And I think things have really, really improved in how organisations think about the impact that their images have on representation. I think your response is a good lead on to the next question, but I really liked what you said about how, we shouldn’t forget about other forms of media.

Nwando
And I 100 agree with you. Sometimes I wonder if that perspective is too old school or too traditional. But hearing you say it gives me a bit of comfort that honestly, I feel like people are so focused on social, socials, which is great, but there’s other forms that we mustn’t also ignore because I feel like if you do it, there’s a whole generation of people that you’d be excluding.

But yeah, thank you so much for that reflection and what you said. And Debbie, David ended with visual imagery, which is really what this whole conversation is kind of bordering around. And so, in your experience so far, how has.

But how has visual imagery really impacted your storytelling? Or how would you say it’s always important to include? Or what’s your perspective?

Debby
Yeah, I think one of the things we found, David and myself from analyzing over a thousand images, used by 32 charities, is that actually in the development sector, humanitarian space, images. And a picture is actually worth a thousand words and more. So that still holds true, although with, generative AI, there’s going to be some challenges to that notion. So, yeah, visual, visual, storytelling is important.

Visual imagery is important in, storytelling. Pictures are important. And it, reminds me of one of the quotes, by Sontag. She says the problem is not that people remember through photographs, but that they remember only the photographs.

And that puts a really different perspective on the kind of images we then choose. Because, And there’s another beautiful, really famous quote by Chimamanda Adichie about, showing people over and over as one thing, you know, and the danger of the single story, or in our case, the danger of the single image, you know.

So, yeah, visual imagery is important. It’s important for so many reasons. It’s important to the work that we do. It’s important for its truth value. It’s important for painting pictures of possibilities. It’s important to document that this, campaign, for instance, actually happens.

But beyond all of that, it’s also having the awareness that these images, are contributing to the public register of global poverty. So it’s contributing to people’s understanding, it’s sort of the educational value of that.

You know, it’s teaching the public what global poverty looks like, the colour of global poverty, the texture, the geographic location of global poverty. So, yeah, in that sense, pictures are really important. As a matter of fact, probably, they carry much more importance than we have given, over the years, or we’ve given thoughts over the years.

So that’s what research, like ours, that’s what it does. It brings back that perspective that this is not just about pictures, it’s not just about fundraising. It’s not even just as simple as telling a story. There’s a whole gamut of considerations to be made, behind every single image. Yeah.

Nwando
Wow. I. One thing I appreciate is the fact that your passion speaks through when, as you speak, David shared an interesting story with me. Corroborates or aligns with what you said about contributing to  the public register of, of whatever it is, whatever topic, whether it’s going poverty, whether it’s, you know, what riches look like, but David, can I come to you, please, withyour perspectives on that same question?

David
Yeah. And I mean, I totally agree with what Debby has said and I found it really interesting. I did a lecture, earlier this week and there were people from all over the world in that lecture.

You know, Vietnam, Japan, Nigeria, Kenya, where I’d just come back from the UK, Pakistan, etc, And I was showing, different examples. This was about social media for development. I was showing different examples, from different NGOs and large charities. And, I played one, which is something positive. And a few of the students were really laughing and I said, why? Why are you laughing? And they were just, oh, it’s okay, it’s okay. And I said, well, no, come on, share with us. And they said it. Why is it always Africa? It’s always Africa. And this particular, film example I was showing them was filmed in Africa. And that’s what came out as well in our research.

So we analyzed, as Debbie says, 1,000 images from newspaper adverts in the UK, and also direct mail. So I subscribed to 10. I donated a small amount of money to 10 different charities to find out what they would send me through the mail. And Debbie and I analyzed those images as well as the newspaper images. 56% of the images, in the newspapers were of Africa and 51 of the images in the direct mail were of Africa.

So those students laughing, they were right. Why is it that all the depictions of need, of poverty, of, well, yeah, people in need, why do the majority of those come from Africa?

And that’s something that we’ve done a number of presentations to large organizations, Debbie and myself, and that’s one of the things that we keep trying to say is “please, please have more diversity in the kind of images that you use. Poverty doesn’t just happen in Africa”. But there was good news as well. One of the things we actually analyzed our data set using a methodology that had been used by fantastic academic Nandita Dogra.

And her book, is now sort of 15 years plus old. And so by actually measuring our data against her data, we saw that there have been some improvements. So, children. Children are always used. There’s always infantilization and feminization in this kind of imagery. But the number of children use had reduced from 42% to, from Nandita’s research to our research, which dropped to 21. So there was a 50% reduction in the use of, of children, and children and women reduced from 72% to 50%. But that’s still 50% of those images, are of women and children.

Where are the whole family units? Where is the diversity? Where are the grandparents? This is something in another piece of research I did, they said it’s actually our grandparents who’ve looked after us. Why aren’t they ever in these photos? So we just need more. Things are improving, definitely, but we need to have more diversity in the images that are produced by Development Organisations, NGOs and charities.

Nwando
Yeah, I agree. I also agree with both of you and, that’s why I really appreciate the fact that you’ve taken time to do this research and to analyze what we’re seeing and to really put it, document it, because if things don’t, don’t get documented, sometimes they just get missed, even though we see them every day.

What you were saying about diversifying images kind of leads me on to what we want to talk about. Well, in addition to everything we’re talking about is – do you think that the use of AI can create an option for diversification of images? Or do you think that could be an option? Do you think it’s a mindset? What are your thoughts around the use of AI really in a nutshell

Debby

David, do you want to go first or do you want me to David, I think you should go first. Oh, well, because yes, I think just a couple of months ago I was at a conference a workshop on artificial intelligence.

Really really brilliant, coming together of academics or academics of researchers, of practitioners, of photographers with experience, working in the global north, working in global south, of you know, humanitarian health organizations.

And it was a really, really brilliant I think it was three days. Really brilliant set of conversations. Thanks to David for, sending me  in his place really.

The whole conversation. There was you know looking at photos in the age of generative AI and there’s a whole lot to be honest I’m now I’m stuck of where to start from because these issues are just flying at me from all the direction.

So yes to your question. Does AI offer …Does generative AI offer a sort of way out to some of the ethical quandaries that humanitarian organisations find themselves in? Yes it does, or yes it seems to do, or yes there’s the promise that it can. But there’s also the pitfalls of that and there’s a lot of things that we need to display awareness on. So it’s possible for, in the case of diversity of cast and characters, you know, the inclusion of more a wider range of people, of the parents, of the men of whole family units as we saw we’re lacking from our research.

It’s easy for this to be generated, you know, for this to be solved in quote, this problem to be, you know, solved by the use of generative AI. We could just sit down and prompt a way and come up with a clean looking campaign. The question however is why do we tell stories in the humanitarian sector? Why do we use pictures? Because the why will always determine the how it will always determine the placement of AI of artificial intelligence in our workflow as communications experts. Otherwise we would be on the slippery slope of doing whatever can be done because it can be done.

So because we can generate family units, we can prompt AI because we can generate, grandparents, the uncles, the aunties. Should we do that? Bear in mind that our research was able to point out, a lot of these critiques because it was based on actual images.

And there’s the danger when we start generating. How do we tell. How can we tell that, you know, what development now looks like on ground, in quotes? How can we tell what the actual representation of participants or beneficiaries are? You know, in the programs, that the charities run, in the intervention. So there’s that danger where we could start generating. We could start making up, and then it starts with images. But at some point will we also generate statistics? Will we also generate data? Will we also generate quotes? Where does the generation stop? Where do we put the mark? And what makes one form of generation? Right. What’s the ethical or the moral consideration behind. What’s the argument for generating images and not generating a quote from an absent voice? You know, so there are all of those ethical considerations, And all of those questions to be asked. And they are not easy questions. They are not black or white questions. They are. There are questions that require organisations to sit down and, you know, and ask why do we tell stories in the first place? Why do we take pictures? And only when we know why can we know how to. Or how not to? So I’m going to take a pause there so I can catch a breath, but there’s a lot I want to still say regarding that.

Nwando
Okay, sorry, David, before you come in, this question of actual representation/ longevity of these images. Take it away.

David
Yeah. I mean, it’s interesting what, Debbie was saying there about, you know, the quotes, because in all honesty, I just hadn’t even thought about that. And so that sort of takes AI generation to a new level. There are so many problems with AI. One of the things that I’ve been looking into recently and there’s actually a film about this that I haven’t been able to watch, but it’s called Humans in the Loop. And it’s about the whole use of data workers often, once again, in countries where they’re using cheap labour.

So Nairobi, where I’ve just come from, there’s. There’s quite a lot of case studies there that people are just literally getting paid very, very little amounts of money to label images to whether that’s on social media platforms with sort of verification.

But it’s also feeding all into this AI system. So there are so many different ethical problems. Because at the end of the day, these images, if you go into Mid journey or Runway or Dali or any of these sort of AI, image platforms, they’re taking images that already exist.

So that’s plagiarism, isn’t it? You’re actually stealing somebody’s image and you’re not crediting them. So there’s loads and loads and loads and loads of different, aspects there. But then there’s also what, you know, AI is improving.

It will improve every, every single day, every minute, probably. But at the moment, there’s, there’s racial biases, there’s gender biases, there’s a whole issue of job displacement. For a start, I, I met with a couple of photographers, while I was in Nairobi that I’ve worked with over the years. They were saying that there are several reasons for this, but, you know, they’ve had hardly any work over the last year. Now they were putting that down to, funding cuts, like USAID funding cuts and things like that.

But I was saying, do you think AI is, has actually, affected your work, the kind of jobs that you do? And they went no. And I said, really? And they were like, no. How would that happen? I said, well, they said, INGOs using AI. And when I said there’s case studies. Amnesty International have been criticized for using an image in Colombia of a protest. And there’s another case, couple of case studies. And they went, ‘really?’ ‘Are they really?’ And they were like, quite shocked. And they were saying, how, how can you use AI imagery? Just that, that, that’s so unethical. And I said, well, it’s happening. So, you know, it’s really, really sad for so many different reasons. But I think one of the problems is to me is these AI sweatshops that not many people are talking about and are unaware of, then you’ve got the whole ethics of actually the amount of electricity that it uses, etc. Etc. Etc.

But then most importantly, when you’re talking about representations, it’s the biases that currently exist within those platforms.

Nwando
Okay, okay. So what I’ve heard so far is, No.

David
Well, I’m going to counter that with another argument. I’m going to counter it and Debbie can continue if you want. We’re talking about the big organisations that can still afford the local photographers. And, and that’s one thing that’s really shifted and something that I think is fantastic is that the large charities and NGOs are using local photographers wherever they work in the, the. Or that it’s increased. They know the local context, they know the local language, they know the culture, etc.

So that’s great, that’s a massive, massive, improvement in the sector. But what about the really small NGOs that exist? The, the community based organisations that literally can never ever afford to use a professional photographer.

Maybe they’re not even using a proper camera, they’re just shooting it on a, on a cheap smartphone. They haven’t got that knowledge and they’re taking photographs. They haven’t got the education. They don’t understand as much about representation.

There’s a really interesting academic paper and I can’t remember the author’s name, from Ghana, who’s explored ethical representation and imagery in Ghana. And basically the answer was that very few people follow any ethical guidelines, very few people have any training and therefore the images they’re producing within Ghana, yes, they’re from local people, but they’re still really, really problematic. So for those smaller NGOs, is AI a solution? I posit that question to Debby.

Nwando
I want to interrupt before Debby goes in. Don’t forget your question, Debby by the way, but as you respond, and I’m going to come back to you as well, David, as you respond, both of you, what about the people on those images?

I mean you’re talking about lovely images, beautiful, you know, and hopefully diverse. But the models, what happens to them? Well, not models, but the people that are in those images. And the reason I asked this question is because many, many years ago, well, not many, many, but my colleague wanted to use AI images on a banner and shout out to Austen, his former colleague. And he argued for it. And I was thinking, I think now, which is like. Or said something similar, which is we have lovely images, why use AI? But anyway, he said what about the people? So Debbie, I leave that on your conscience.

Debby
Okay. So yeah, it gets complicated the more you dig in, the messier it gets. And that’s just and that’s just the thing with AI, with generative AI, with technology anyway, as it were, but especially with AI and much more so because the speed at which things are changing is alarming.So, which is why this is just an aside anyway, which is why just developing policies or guardrails won’t suffice. Because there will always be new use cases, there will always be, you know, new advancements and that your policy doesn’t necessarily account for.

So, yeah, having said that, back to the question. AI for smaller NGOs and the people in the picture. And before I just come to that, I really wanted to speak to the point David was mentioning about, because Amnesty International were in that room, in the conference organised by Dr. Aseni and Dr. Sonia, regarding the use of artificial images. And Amnesty International was there. No MSF. Rather, MSF was there. And we had that conversation about using generative AI, large organisations using generative AI.

And there was also the question of carbon, decarbon. What’s the word? Decarbonisation. I don’t know. Reduction in carbon emission. You know, which is one of the things that led to or one of the arguments behind hiring local photographers in the first place.

You know, so rather than having somebody from the Global North fly into Global South, burning thousands of energy, and oil and all of that and pollution, just hire somebody locally. That being said, I find that the argument for the climate sorts of balances itself or sorts of. It goes both ways. So while we’re saving on fuel, as it were, on energy as it were, on one side of the argument, using AI does not necessarily prove to be the most sustainable, use of Earth’s resources either.

That’s on one side. Again, I have digressed because there’s a lot of thought in my head. David talked about, the use of local photographers. And I worry that with generative AI, all the grounds we’ve sort of gained, in terms of localisation, in terms of shifting power will be lost. And this is because I don’t see that we gained so much ground. Anyway, we need to appreciate that there was progress being made, there was progress being made in terms of hiring of local photographers. But from that conference I was at in Geneva, the photographers in the room, had to say that the briefs were still a barrier.

So yes, we were hiring local photographers, but we were sending briefs from the global north to the global South. And now in place of that for generative AI, the briefs have now become the prompts. So now, who gets to decide what an image looks like?

Who gets to decide what is being generated. So, we talked about co-creation in previous spaces. So now we’re going to talk about co-creation of prompts. Are we going to get people with lived experience to actually dictate what an image should look like, would look like? You know, so there’s all that conversation around power, and around relocating power to the global north or to large organizations or to the organizations using artificial intelligence, AI generated imagery. And so AI for smaller organisations. I wouldn’t say yes or no, David. I don’t know because it’s not that straightforward either. But I do think it holds promise. I think it holds promise, because it’s accurate to say that some organisations do not have the budget or the time or the luxury of either of these for hiring local photographers, for working with local photographers, or even global photographers as it were.The consideration then would be the truth value and the trust value. So for local organisations using artificially generated images, would they pass this off as real images taken by real photographers? Or would they be willing to label this?

Would they be willing to disclose this? Would there be that transparency, that accountability? Now there’s been some research pushing back against NGOs using AI and I think it’s because it’s still early days. I think several years down the line the resistance will grow, less and less. There will be backlashes. But I think that with time, the donor publics will become more accepting of that. But before we get there, we don’t want, what we don’t want to do is to be in too much of a rush. There’s a lot of talk about adoption.

There’s a lot of organizations adopting AI rapidly. But we cannot force the donor publics unless they are ready for that. If we do so, what the sector would lose is trust, the erosion of trust. And that’s a really costly currency for humanitarian organizations.

And then to question about the people in the picture. So, one of the considerations, one of the controversies, or one of the challenges with using, with taking images of real people, has been the issue of consent. How long does consent last? And then how well can these images be repurposed into other images, or into other formats? How do you reuse these images? And then how about the procedures for recall? So, where someone is no longer interested in having their images in the public domain, how do you recall that?

Well, you might be able to recall the post, you might be able to retrieve, you know, what has gone Out. But how about what is already, stored in the memories of, of audiences, global audiences that already received these images. How do you deal with recall?

So in that sense, it seems that generative AI offers a way out. After all. We can argue these are not real people. These are this person doesn’t exist anywhere. You know, but again, there’s no easy way out. There’s no easy way out.

That also has, that also has its challenges. I think I’ve spoken quite a lot, so I’m going to take a breath now.

Nwando
Yeah, that’s fine. I’ve thoroughly enjoyed, your reflections and I like the fact that we’re not necessarily seeking a solution. We’re just thinking through all of the different, likely solutions that exist and really talking through the conundrum as comm people who are now dealing with this, you know, this change and trying to really navigate it in a way that we’re not harming anyone. David, would you like to comment on ‘what about the people?’

David
Yeah, yeah. I mean I’m gonna quote Debbie here because I think that’s funny and I’m going to use it in, in lectures and the briefs are. But it also made me think while Debbie was talking about, somebody who used to work for a very, very large charity, I won’t name them.

And they were in a motorway service station in the UK and they went to use the, the washroom. And on the back of the toilet door was actually a poster from a charity. And they laughed at themselves and they thought, I wonder when they actually consented to the use of their image, did they imagine themselves on the back of a toilet door in a service station in the, on the outskirts of London?

And, and it’s always made me smile that because I’m thinking, no, they did not. They had no understanding. And however many ethical guidelines you write, however much training you use, do the people who are having their pictures taken, do they really, really fully understand how their images are going to be used?

And is there a process in place to show them, to go back and show them how their images have been used. So to that extent, AI is a solution. And I, I read something on the BBC the other day and at the bottom it said this article has been. I can’t remember how it said exactly, but rewritten using AI. And I thought, well okay, so at least they’re being open and honest. The, the thing that I always ask the question because I, I teach in the school of global development, but also the business school. And it’s interesting having those conversations with different people. Everybody will, you know, when you ask the question, is it better to use positive image, negative, image, no image at all, or if you add in now an AI image, everyone’s going to have a different perspective, on what is ethical.

Because to some people representation is going to be the number one goal. To others, they consider the fundraising is the number one goal and there is no right or wrong answer. And the people that are desperately in need, they would probably say that fundraising would be the goal. Whereas if you speak to elites in different countries, they’re saying it’s representation, the diaspora representation, the academics representation. What about the people really, really suffering to survive, literally to eat, to survive, to have access to clean water, etc, they might have a different perspective. So they might say it doesn’t matter, I just want the most money.

Debby
Absolutely, absolutely. And I think just to, just to jump on that, David. So I give quite a lot of thought to ethical storytelling or ethics as it were. And how ethics takes a different shape or a different you know, texture across cultures, across space, across time.

And what we consider ethical, today might not necessarily be so. And then we, if we try to judge the ethics of the past with the eye of the now, you know, we would be horrified at some of the decisions that were taking. You know, so yes, there’s the need for that humility, that understanding that we are humanity is flawed where imperfect humans work, making the most of and the best of an imperfect situation.

And it’s that awareness that we haven’t got this right. We probably will not get this right on this side of the divide, of this side of eternity. But we can try to do the very best that we can, while we can, by centring people, people, over technology, people over principles. Yeah, people, people, people over technology. Absolutely. And people over principles. So yeah, those are some of the ethical conundrums. And there’s also the challenge then of organizations in the global south because when we talk about ethical storytelling, there’s the tendency to narrow it down to Global North organisations, you know, humanitarian organisations situated in the Global North.

But does ethics also apply to organisations in the global South? Does it also apply to local organisations? Does it also apply to my local homelessness charity here in Eastbourne? Should they be worried? Should they be asking the same questions, otherwise are we being pretentious about it? And that reminds me of, you know, brings me back to David’s thoughts about ethics takes a different shape or taste depending on who you ask.

David
Can I just add something because you’d reminded me. So I just looked up while Debbie was speaking that the, the academic paper that I was talking about in Ghana. So it’s by Mahmoud and it was published in 2024. So as I said, that looks at NGOs, 22 NGOs, and it looks at sort of their storytelling, their ethics. And this, this whole sort of pornography of poverty debate.

So that, that’s a really, really interesting paper and it’s great, as Debbie said, it’s great that these discussions are now going on in the in the global south as well as the global north, because it’s important, because there have been shifts in fundraising as I saw in Nairobi, with people from SOS, raising money there. So these issues need to be discussed globally, not just in the UK and Europe.

Nwando
Absolutely, I agree with you. And the good thing is HLA’s listeners are quite a broad sphere so hopefully we’ll get some reactions to how people are thinking about this from all areas of the world.

What would you say if you were to give steps, what would you say people should be doing? And this is not necessarily recommendation because I said before, it’s light-hearted. It’s not like we’re saying, oh, you must do this. No, it’s just in your own experience – what would you say people should be thinking about, be cautious of if they are going to employ AI for visual storytelling as humanitarians?

David
 Okay, do you want me to go first, Debby, or do you want to,

Debby
Okay, let me go first. Let me take a quick, swipe at that and again I will. You know, so when, when you ask a new question, Nwando, my mind goes back first. It’s just like a rubber band. It pulls back first to what we’ve discussed before. It snaps forward, you know, so, and then one of the, with regards to your question about whether this is a yes or a no, and of course we know it is neither or it is both.

There’s also the issue of gaining efficiencies with regards to the amount of, of data being demanded by humanitarian organizations. So on one hand there’s shrinking budgets, as we’ve seen. There’s also the pressure to keep up on social media.

How many posts, how many more posts do we have to make these days now than before? How many more? What’s the lifespan of a story today? What’s the lifespan of an image? How often do we, you know, need to tell new stories in new formats to new audiences on new platforms?

You know, and how quickly can we go on, data gathering, dream trips, on you know, verification trips and all of that? You know, so there’s that, there’s that demand side of storytelling, which is, you know, is driven by changes in media.

We’re all being mediatized one way or the other. So there’s that gaining efficiency. So, AI in that sense sorts of provides an escape route, where, or it gives promise in terms of scale, in terms of output, in terms of speed, that does not necessarily make it right or wrong.

But then I’m coming to your question. If an, if an organization, if a humanitarian organisation, a charity, someone working in the third sector decides to go the AI way, which to some extent we all will. I think there are just a few considerations that pop up for me.

And first of all, it’s to, to really decide as an organisation what is most important to us. What’s our mission? What’s our, what’s our mission that must be up and central what our values as an organisation because how we, I mean, how we tell stories will change over time.

There will be new tools, there would be. Let’s not even think that AI artificial intelligence would be the apex of technological advancements. Once upon a time we were debating whether to use smartphones or to use analog cameras. Once upon a time we’re debating whether to go natural or to you know, to use makeup or to, you know, to augment those sorts.

So there will always be constant advancements. There will always be. Humanity will continue to push forward, push forward in tech, in tools, the medium. You know, so when we know the how, I mean, when we know the why, rather, then we can figure out the how, you know. So, yes, that’s one of the very first things for me. It’s really defining, really understanding, really centring what the values of the organisation is. So before you dive in, you know, what, what are we really about? What is our vision?

What is our mission? And where does artificial intelligence, where does artificially generated imagery, where does it sit within our workflow? You know, and it’s also important to note that the change should not, should not be built around a person or a trend or even a new technology, you know, change.

So what often is the case is we have one person in the team driving or championing the cost of adoption. One person saying, let’s adopt this. And that often happens with a lot of issues like diversity, equity, inclusion, decolonization. That is such a mouthful. Like, you know, there’s always one person saying, oh, let’s get on board, let’s do this. But change cannot be sustainably built around the person. It has to be organisational culture. It has to be embedded in the culture of the organisation.

The enemy here is the time it takes to do such a solid groundwork. You know, things are moving at breakneck speed. We don’t have the time to sit down and ask deep, tough questions. And this is where, you, you could get external help you could leverage on people like David and myself.

This is not free marketing, I promise you. We really want this stuff. So yeah, this is where you could leverage an external help to facilitate deep, quick sessions. You know, you could get the relevant stakeholders in a room and you know, get this conversation started, get a blueprint, get a strategy in, in place.

Because it’s not about developing clear policies or currently if you just take a quick Google, I mean if you search on Google, you will find a lot of AI policies, ethics of AI Usage, guardrails, that’s great, but it often gets forgotten in use. What gets remembered is culture. It’s culture. So before a list of do’s and don’ts, a clear sense of why is better. So, yeah, I mean, those are some of the things I would say. You know, and rather than making broad sweeping changes, rather than all of a sudden tomorrow we’re back in the office and the entire comms has been replaced by an assemble of AI agents.

Rather than making broad sweeping changes or general, adoption, it’s good to ask where can we fit AI into the flow? It’s always the technology should be solving a problem. It should not be leading the way. We should not be adopting for the sake of adoption.

Where does it genuinely help? What are the efficiencies to be gained? How can we test, how can we track and how can we iterate? And finally, I think the last point I would make is that while adopting, artificially generated use of artificially generated images, we must display an acute awareness as a sector.

When we started to put forth images for fundraising, in the earlier days, little did we know that we were creating lesson notes, we are creating textbooks. You know, we were writing the textbooks that AI would be trained upon. We didn’t know then, you know, as there was not that, that awareness, there was not that foresight, there was not that thinking, that the images we were putting forth had any significance beyond just raising money.

But now we know better. Dr produced some really brilliant work about, the biases in artificial intelligence and how AI has taken our bad lesson notes and has run with it and is multiplying that. So when we are adopting with generative AI, we must just display that higher level of awareness of our role as educators.

Even with the use of generative AI, even with generated images, we are still educating the general public on what poverty looks like on, you know, we’re still educating the general public on people, on, places. So we must display that awareness. Of course, we don’t know 10 years down the line, even tomorrow, we don’t know what the repercussions of some of these might be.

But we can just, just, you know, make the best with what we know in, with the imperfect systems, with the imperfect man and, and forgive ourselves and, and move on. Absolutely. So, yeah, those are some of the, a few things I would, I would say not much of a step per se, but yeah, more like questions to be thinking about.

Nwando

Yeah, absolutely. Useful. Thank you so much. David, what’s here?

David
Well, as Debbie has said so much, I think that was probably about a 17 minute answer. I’ll try, I’ll get mine down to 17 seconds. I would say be, be transparent if you’re gonna use it, because that’s important for trust. Make, sure you’re authentic. And as Debbie said, you can have as many guidelines as you want in the world, but how do you implement them? And how do you monitor them? You need to implement guidelines lines in order to be successful. So there you go. Was that 17 seconds? (laughs)

Nwando
I think those suggestions have been so, so they’re so apt because I feel like we’re very, as a people in general. You know, I remember when you mentioned, smartphone, there was a time when people were scared to use social media. I remember my roles. I was trying to, I was literally begging the senior management to please join Twitter and Facebook.

They were like, no, their privacy didn’t. I was like, trust me, nobody’s going to scam you. Because it was just, I feel like social media came just after, you know, the scamming of, oh, I’m a Nigerian prince here to, you know, I did not come to you by chance. I feel like that was when like Facebook started just after that whole thing of email.

You know, we’re still, we’re still trying to get over email. And then, you know, we got Facebook and Twitter and all. And I remember there was so much restraints, especially because then I was working in Nigeria, so there was so much, you know, people were like, I don’t want to mess about with this. And now we’ve got now AI, so things will keep on changing like you said. And I think that all of these points that you raised, I really like about being transparent. I really, really like that. And of course, know your why it will always be because that leads to authenticity. If you don’t even know where you’re, why you’re doing what you’re doing, how are you going to be authentic about it?

So you guys, you guys are just superstars. Thank you. But lastly, lastly. Lastly. Well, ish. I wanted to ask the question about prompts because it was mentioned earlier in the conversation because it links to briefs as well.

And this also links to previous questions like what, what should we be thinking about? And if people are deciding to do all of these, I know that, that their why would help them be a bit more, you know, hopefully help them to do no harm. However, there’s a lot of lack Of, Well, I don’t, Well, I don’t say lack of information, but there’s a lot of people don’t really know.

Everyone is like, you know, experimenting. So, what, what should people be thinking about when it comes to problems? And this question comes with, from the angle of localisation which you mentioned earlier, and really thinking about telling people, telling people where poverty or where riches or where, you know, different topics live.

So yeah, sorry if that’s not clear and feel free to ask me what, what are you actually asking?

David|
I, I think it is clear. I would, I would say I, I loved what Debbie said earlier about co-creation. I think if that is possible, then that should really, really, really be explored to ensure, localisation, context, etc.And then I think don’t be lazy with your prompts. You have to be really detailed. You can’t just type in, in your first instance what you want, you have to keep adding to it. And if it’s not right, you have to change it again and change it again.

But as, as to who needs to be represented, I hope that the, you know that the charity advertising website that Debbie and I have set up has loads and loads and loads of images as examples. It has a lot of data analysis.

And I would say that the main findings are that the diversity of characters. Stop, stop always, using Africa as an example. But equally if you’re working in Africa then you don’t want to, you don’t want to include images of India.

Somebody actually wrote to me and asked me that question, said but I only work in Kenya, what can I do? And I went, you need to make images from Kenya. You know, just make sure that you try and include the whole family units. But yeah, lastly, because I hadn’t really thought about this in those prompts wants make, try and make sure that you include those kind of details, those, you know, what is missing from the imagery at the moment.

It’s all family units. It’s fathers with their children, it’s grandparents with their children, it’s grandparents alone, etc, doing different things. So what’s missing? What could actually be a much more accurate representation of what you’re trying to portray?

Nwando
There was something you shared with me earlier David, about the evolution of that, you know, of prompts and how you educated – you’ve been taking time to educate, which I think is, is a very valuable thing to do.

Is if you see, just like you said now, if you see something missing when you do a prompt is keep, keep, keep, keep reiterating, keep saying no, add this, add this at that. And then with time, hopefully we all, get to the point where, images are. Well, AI is making images that we can use if we decide to use them. Exactly. Yeah. And, I think, I think, I feel like we can pause this conversation here, but if there’s anything else that we feel like we’re missing out on

Debby
Yeah, I think just to, Because I like this question about prompt and I’m all of a sudden thinking I’m going to be in David’s DM afterwards, we should actually run a course on how to prompt better, you know, how to prompt effectively with all of these ideas, of localization, of shifting power, of access, of doing no harm, of our roles as educators, as charities. But yes, educating the public. So I think it’s something, it’s something to be considered. And the point I wanted to make lastly is that, like David said, don’t be lazy. We’re going to get to a point where AI keeps improving, the images being produced keeps improving, but it’s always a matter of garbage in, garbage out.

So if you don’t have the background knowledge of what a good picture looks like, or what a good picture should be like, or what an effective picture should be like, or what a picture should not be like, if you don’t have all of the knowledge in you, know of the critiques we’ve had over the years, then it’s really easy to do a lazy job, a really sleazy job.

So I think one of my recommendations would be, to go and dig into the charity advertising website, read both, reports, dig into that, look at what’s been done so far, look at the kind of images, look at the loopholes, look at the gaps and then, within your context, within your own reality, within the work that you do, your operational reality, be creative. And yes, absolutely, for transparency and authenticity.

Nwando
Thank you so much. And that research will be linked in this, in the podcast page when we publish, as well as other resources. I found David’s blog as well. So all of these resources will be linked for our listeners to engage with. So on this note, I want to say a massive thank you to David and Debby. Thank you so much for sharing your wealth of experience with us this, on this episode. And I honestly, personally, I genuinely appreciate everything that you shared.

I think it should be very useful not just for comms professionals, to be honest, but for everyone who’s thinking about navigating the use of AI in the humanitarian sector. And so thank you again for your time.  We look forward to that course.

Debby
Yes, absolutely. We’ll send you a discount code. Honestly, only. 999 only. Only if you. If you sign up now. If you sign up now.

Nwando
And, consultancies, I think after the course, I think that doing consultancies one on ones with organisations who are really trying to figure it out because this is not the easiest thing to figure out.

Debby
No, it’s not. Yeah. So you need help.

Nwando
Yep. We’ll reach out to you guys. Thank you so much, Deb and David, and, thank you for joining us on this episode.