How to Win with AI, Part Two

This episode of The Whole Idea podcast by DCG ONE is the second of a two-part series on artificial intelligence. Designers, a creative strategist, and a data analyst from the Agency at DCG ONE share their thoughts on how they use AI to make their days more productive and projects more successful. We also spend time chatting about ways to make sure AI technology evolves properly and safely. Join host Greg Oberst for this timely discussion about AI. See what IA can do to help you streamline projects, save money, and free up more time for what you do best.  

Other links you may like to check out:

About us - https://www.dcgone.com/about
Strategy - https://www.dcgone.com/strategy
Technology - https://www.dcgone.com/technology
The Agency - https://www.dcgone.com/agency
Let's connect! https://www.dcgone.com/contact

Email us: podcast@dcgone.com

Check us out on social media:
LinkedIN, Instagram, Facebook

Transcript

Greg: Hello again and welcome to the Whole Idea Podcast by DCG ONE. I'm your host, Greg Oberst, senior writer at The Agency at DCG ONE. Thanks for tuning in. This episode is the second of our two-part series on artificial intelligence. In part one, we heard from DCG ONE President and CEO, Brad Clark and DCG ONE, Senior Vice President of Technology, Chris Geiser. Then, we explored where the best opportunities in AI exist for businesses. We also looked at how to manage the daunting security and privacy issues that come with AI. Some good stuff in that episode. Check it out if you haven't already.  

In this part two, we'll chat about how artificial intelligence tools are applied at the agency at DCG ONE and get some additional thoughts about AI that are on the minds of our guests. And there are quite a few thoughts there to be sure. Let's start with our Director of Creative Strategy, Shya Scanlan.  

Shya, just what is artificial intelligence?  

Shya: I have no idea.  

Greg: Alrighty, then. Thanks for coming.  

Shya: Honestly, I mean, there is no part of me that would hazard a cogent explanation of what artificial intelligence means.  

Greg: Not too many people can.  

Shya: AI is complicated, and I think that even those who are most schooled in it readily admit that there are limitations to their understanding of what it does or how it does it. So that's just startling and profound on its face. Anyone who's dabbled with AI in its many forms, whether it's (large) language models or on the visual side of things, are just simply flabbergasted by what it can do and how quickly it can do it. And that is cause for some excitement, which I feel and also terrifying, which I am on occasion.  

It is again, as I said, profound in how quickly it can create something seemingly out of nothing, which anyone in the creative field is asked to do almost daily. So it basically is, it has an easy time doing something that is a challenge for the brightest and best of us. That's not to say what it is creating is always of high quality, but it is startlingly high enough quality to feel like it could be a threat.  

Greg: So then how intelligent is AI? Sometimes it can come off kind of dumb.  

Shya: Well if you think of what intelligence is generally, or if you were to ask what sorts of things do intelligence or intelligent people do, I think that you quickly encounter some pretty compelling overlaps. One that comes to mind immediately is synthesis. Intelligent people are able to create novel synthesis between disparate ideas or sources of information. They see patterns. And one of the things that large language models are good at is that same process of scanning huge amounts of information, in fact far more than any particular individual has access to at any given time, and finding patterns and synthesizing that and reflecting that synthesis in a way that is consumable and intelligible to human beings.  

So I think that is pretty amazing and also pretty understandable or relatable. The fact that it does that makes it relatable to me as a human, I can say, oh, that is something that you're doing or that this engine is doing that makes sense to me and that I can immediately think of use for again.  

Shya: And that's what probably makes it a little bit scary to people is that it's repeating not just something that's abstract and meaningful only to specialists, but something that is immediately comprehensible to the average person. So it begins to feel person in that regard. It's cool and scary, but it's something that a lot of what I do in my job is synthesis in creative strategy.  

In all strategy, you're surveying a landscape of information from a variety of sources and you are trying to create, you're trying to find uncover patterns. And from those patterns extrapolate some kind of actionable information or perspective once, especially as these language models are connected to the internet and so the sources of information are live, that becomes a really powerful tool. It can be directed to survey a competitive landscape for a client in a particular industry and find what sort of patterns happen to be occurring.  

Shya: What are people normally saying? What do all the top 10 companies in category X, what do they look like? What do they sound like? Do they say something similar? What claims are they bringing to the table? And by establishing that information, which is something that we do with our clients all the time, but by being able to do that at least as a baseline using AI, it brings us that much closer, that much more quickly to the finish line of a recommendation, for instance, or a set of potential directions to head creatively or strategically.  

We hear a lot about prompt engineering to be able to converse with a language model almost and to see what it's generating and to then go back and ask it to refine those generations or those outputs. There's an artistry to it. It's not just a science. And I think that that's alluring, that's compelling, that makes me want to engage with it. I don't think it's an accident that so much content that's generated with AI tools is art or creative. It sparks that within us. It asks us to create further and to dream again on its face. That's exciting. It's not just for engineers, it's for poets and artists and musicians and singers. And so that's the kind of way that I look forward to using AI in my work.  

Greg: That's Shya Scanlon, our Director of Creative Strategy. Next up we have Kristen Crandall, the Director of Analytics here at The Agency. Kristen, maybe you can define artificial intelligence.  

Kristen: Well, two things really. I think AI in general is kind of an umbrella term. Now, that's loosely any machine or software or algorithm that is solving a problem that's usually done by a human should be able to learn as it goes and mimic our own cognition in a lot of ways. That said, I think what most people think they mean nowadays when they say AI is a learning language model, that's what ChatGPT is. And that really is just a program that's guessing what the next word is going to be.  

So there's an image that always comes to my mind when I think about this, and it's an old Saturday night live sketch with Kristen Wig and Fred Arman, and they used to show up to a weekend update unprepared and they were like a singing duo. And so they'd improv and they'd stare at each other and guess what they were going to say and then sing a song on the fly. And I think that that's how learning language models work.  

Greg: Okay, that's a little scary. So Kristen, let's look now at how AI can help our business and our clients.  

Kristen: I think the answer to both is kind of the same, but it helps us first hopefully be more efficient, which goes back to taking some of the tasks that just take time, but don't take a huge amount of thought. So taking some of those and finding ways to automate them makes us more efficient, makes our clients more efficient. I think that's the first thing. And then the second is hopefully it'll make us faster.  

So even with some of the tasks that are maybe one step removed from basic or even the basic tasks, they can take a while to do right, hours, a while, hopefully eventually AI can do that in minutes instead. And then the clients, and we are getting to that easy data really fast, hopefully accurately. And then we spend time just thinking more, which is such a luxury to have time to think about the data. But I want us as analysts to spend more time thinking about it. That means the insights we're giving our clients are better, and then also our clients can spend more time looking at those insights instead of trying to find a bunch of signals in it and do the same thing, try to really keep digging deeper and take action.  

Greg: Earlier, Kristen, I was detecting a tad bit of cynicism. What's behind that?  

Kristen: Well, there's a couple things. One is just I've seen some output that doesn't look like what I would accept. Even if I had a business analyst delivering something to me that was brand new, I kind of look at it that way and it's just not up to snuff in terms of accuracy for me.  

The other thing is it's trained on the internet and the internet is full of garbage. And I also know that what they're currently trained to do is give you an answer. Their main goal, their reason to be is to give you an answer. And it doesn't really matter if it's right, it just matters. Like here's an answer that seems like maybe it's right. So I don't trust that it's trying to be correct. And so for that reason, I'm super skeptical. And then I also think it's happening really fast, which isn't a bad thing. It's great to innovate and evolve, but I just don't see evidence of a lot of methodical, careful thought, which we have to have as an analyst to make sure that it's trustworthy.  

Greg: Yeah, there are certain, let's call 'em philosophical issues that do come up with AI. What are yours?  

Kristen: In my profession, it's accuracy again. And then that we'll come to a place where we'll start accepting a certain level of inaccuracy because it's so much faster and simpler. I really, really hope that doesn't happen. And then both in my position and just in the world, I'm afraid that it is going to impact our ability to think cognitively.  

When I think about being in college and how much time I spent with actual academic textbooks open in front of me and hours and hours and hours trying to solve a problem or get the right context even though it was painful, I'm grateful for that time because it taught me how to slow down and to problem solve. But if I don't have to do that on my own, if I can just ask a robot to do it, then we're building only artificial neural pathways. What happens to the actual neural pathways in our brain that problem solve? Are they just going to grow over? Are we going to lose our creativity? Are we going to use our curiosity? Because you can get an answer immediately. Those are my big philosophical worries, I think.  

Greg: That's Kristen Crandall, director of analytics at DG one. Next I caught up with Justin Parnell from our New York office. Justin is an Associate Creative Director. I asked Justin from a design perspective how he was seeing AI tools come into play at the agency.  

Justin: Yeah, it's been interesting for myself and the design team, how we're just kind of sniffing around all the various tools that have been baked into a bunch of our software that we're using day to day. It certainly will and is starting to play a central role in how our team works here at the agency. And it's kind of helping them, at least from the initial get-go and experimentation, getting us to a first draft or a concept faster, and then designing and building a great new experience for our clients from there.  

So again, a lot of it is just similar to that wandering of new imagery or a copy that can be pulled in from AI just to explain ourselves in a conceptual part of the phase of presenting to a client at that point. And then I think in that discovery phase as we call it here, we're exploring AI that might help us generate or synthesize early ideas with a single prompt.  

Justin: And I think we're all kind of getting used to ChatGPT asking a question. It might come up with some general ideas, whether it's obviously a copy or even on some scripts or even some design ideas that could maybe prompt us to go to a direction visually. And then I think a lot of times, even when we're coming back into the design, again, it's just using AI to tap into our existing designs or even our existing design systems and then really even surfacing components that they could recommend to us to kind of build upon a new experience, whether it's obviously a new website or a mobile app.  

But again, I think the more we're able to look to AI to make our lives more efficient, I think this is going to save everyone headache. It's going to save a lot of money. It's even saved the clients’ money because now we're able to really speed through and create things for them in a faster, more efficient way.  

Greg: Okay. I heard ChatGPT in there. What other AI powered tools are the design teams using?  

Justin: Yeah, so I think Dolly is the other big one that our design team has really started to leverage. It's a little bit on the more visual side to come up with some very unique visual systems. Maybe it's an illustration, maybe it's a photograph to really kind of showcase what we can do from a conceptual idea for a client rather than say it. And I think there's even instances where we've been able to utilize this type of tool to showcase where are we bringing in new patterns or textures to a visual identity to a brand that we're onboarding here at the agency.  

We also have other tools that every designer uses within the creative cloud, the Adobe Creative Cloud or Figma. And all of these are starting to have just AI kind of built within them. I think the big thing that I think we're all celebrating, especially in Photoshop, is AI's ability to remove backgrounds from an image.  

Justin: So historically, this took hours, and it was almost one of the biggest nightmares for most designers, especially production designers, to spend hours and hours basically clipping an object out to remove the background. So if you think of someone with a lot of hair or even pets, you really had to go zoom in and just start masking them out, tracing them out to remove that background. AI and all these other tools have basically done this in a one click prompt.  

So it's pretty amazing to use these tools and I think that's the power of AI is to make a lot of this monotonous stuff kind of go away. It's literally allowing us to really kind of focus back on the actual design and finding solutions. And again, just speeding up our entire process more and more.  

Greg: Are you worried at all, Justin, that AI will be assigned to execute too much of the creative?  

Justin: Yeah, I hear you. And having this fear of what AI can, so-called replace us. I actually don't see it as a replacement by any means. Again, I think it's a way to use it as a tool to sort of maybe get some ideas flowing. And I think for writers like yourself, I think we're always going to have to learn to edit what is our tone, and that will never be replaced with AI and ChatGPT.  

It's that true human connection that is through our personalities that really can't be shifted or translated through AI from what I see it. So for designers, it's the same thing. I think we all have a very unique style and perspective that a lot of it is influenced by our own surroundings. A lot of it comes from the clients say, brand guidelines, and that's not something that I see so far that can just be entered into an AI prompt and then sped out. 

Justin: I think we're still, as designers, we're still going to have to design a design system. And when I say a design system, this has been around for a while, but it's probably been the most celebrated thing for designers because it makes us more efficient and we're not having to focus on say the exact radius of a button, and it just keeps things a little bit more concise and consistent. But I think AI is going to be the biggest consumer of our design system. So it could take all of these components from our library and our design system and then perhaps produce a new experience. But I still think a designer is still going to have to go in there and adjust and become an editor to make it truly unique for that client. So again, I don't think in my optimism, I don't see it replacing us by any means.  

Greg: Justin, I'm thinking that technology as powerful as AI has the potential to be, should probably be approached with some measure of caution or concern. Do you have any concerns?  

Justin: I think there's a little bit of concern I've heard from other designers and maybe myself, is I think AI is actually going to start to evolve roles a little bit more. I think one of the, I guess more important pieces of this puzzle is how designer roles and collaboration might change in the age of AI. So at The Agency at DCG, we need to understand and believe that personas and workflows will shift, and we need to be open-minded to that and not be scared to that.  

I think more people will become visual creators and writers because of it, and existing designers will be able to create more ambitious experiences than ever before. I think one thing to maybe also think about it is the total space of design having a ceiling and a floor. So in this instance, I would say the ceiling is how a good designer can be at designing, which is often constrained by the availability of whatever tools is available to them, what software is available to them.  

Justin: The floor is the minimum skill required for someone to participate in design. So this often pushes out, say people over in accounting or a project manager to become a quote designer. I think AI is going to allow us to lift this ceiling leading us to become more of a, I guess leading to more creative outputs made possible with all these new tools. And we'll also lower the floor, making it easier for anyone to come into the design process and visually collaborate. So this excites me actually. It doesn't really scare me that we're replacing people's role. It just allows everyone to finally collaborate and work towards a common goal and have this shared responsibility.  

And I think that's something truly unique to our agency of particular size, that we want to hear other people's experiences, but oftentimes they feel limited because they don't have the tools to say right or design. But if they have tools within AI, it allows 'em to get that visual expression that might be locked in their head. And then it's up to us as the design team to kind of push that into a new visual experience. So again, I don't know if I would ever call it a fear of technology. I just find it interesting that some designers are saying everyone now gets to be a designer and that scares 'em. And I think we should just embrace collaboration a little bit more because of these tools.  

Greg: Interesting perspective from Justin Parnell, DCG ONE Associate Creative Director. Finally, Robin Harrison. She's an interaction and UX designer at The Agency whose work with AI powered tools is similar to that of Justin's. But as we spoke, Robin offered an interesting perspective to my question about how to control AI and safely work it into the mainstream of our existence.  

Robin:Yeah, that's a great question. I mean, I would say because we know that AI is feeding off of data, that we feed it as human beings and anyone working in any field where they're able to interact with AI, we need to always be feeding it positive information. I think that the worldwide web is a dark place sometimes, and if it is scraping the internet, we know that if we look at sites like Wikipedia, Reddit, Twitter is kind of like a wild west nowadays. There is some pretty dark content on those places and that sometimes if you ask an AI tool to write a story about a woman or a person of color, it's not going to do the best job because it's going to have bias and racism based on what it's seeing online.  

So I think it's super important for us to empower each other and really make a space intentionally for people of color, LGBTQIA+ people, people from different cultures to have inputs into that AI tool and to get their own lived experiences included in that so that we can build as a UX designer, we need to include all of the populations so that we can get a very fair result I think in terms of what we're getting back for prompts and for storylines and for all of that with AI tools.  

Greg: Robin, are you one who believes that some level of oversight is important to make sure that accessibility and representation are there and are accurate?  

Robin: I do. I absolutely do. I mean, I do believe in the concept of safe spaces on the web, if you will. So what I mean by that is we know Facebook now has an entire department devoted to taking down hate speech or hate videos online. Man, that is a rough job. I do not envy those people. I mean, I've read stories about the individuals who do that for their living, and I can't even imagine the things that they see and read every day. And I think that now that sites like X, formerly Twitter, we know that there's a lot of antisemitism and hate happening on those sites. And again, this is all internet, all data that these AI tools are scraping. I do think it's important to, I guess, create the world that we want to live in and that we want to be a part of.  

Robin: And part of that is if we are teaching AI as a tool about humanity, then again, we need to represent what we want to see in humanity. So we need to teach it to not predict these biases or create more discrimination. I mean, the last thing we want to do is create more harm. And I think as a UX designer, we're always thinking about the products that we're building.  

I personally am very fascinated with unintended consequences in UX design. And I think one of the things that we are always asking ourselves, especially here at the agency when we're thinking about building a new product, is how can this product cause harm? So I think we need to always be asking ourselves that with AI, when we're thinking about how people are building it, how we're feeding information as humans, how could this potentially cause harm and how can we mitigate or ideally completely remove that harm?  

Greg: Thank you, Robin.  

Robin: Thanks, Greg.  

Greg: Some final thoughts from me about AI and what we heard from this team.  

Clearly there's shared excitement about what AI tools can do to spark creative thinking, streamline projects, generate more data faster so that strategy and decision-making can be better informed, not to mention opportunities to be more efficient with budgets and even save money for our business and for our clients at the agency. I feel like we're on top of that.  

I was also encouraged by the thinking and awareness of AI's youthfulness, so to speak. It seems like AI is a bit like a child with infinite potential, and like a child, AI needs guidance to grow properly and governance to make sure it continues to be a resource for good. 

Thanks again to Robin Harrison, Justin Parnell, Kristen Crandall, and Shya Scanlan for their thoughts on artificial intelligence in this episode of The Whole Idea Podcast. If you have questions for our guests or about anything you've heard here, write us at podcast@dcgone.com.  

Thanks very much for listening, Whole Idea Podcast producers are Mandy DiCesare and Kelcie Brewer. I'm Greg Oberst. Watch this channel for our next podcast and more expertise, insight, and inspiration for Whole Idea Marketing.  

Take care.