Chris Duffey is an expert on how Artificial Intelligence is enhancing creativity today. Chris, a MIT trained, three-time Cannes Lions speaker, has worked as a creative for over twenty years; most recently as Executive Creative Director at WPP. He is now Senior Strategic Development Manager at Adobe.
Whether you work as an advertising creative, as a director, an illustrator, a photographer — or across almost any medium throughout the arts — AI is already able to enhance your creative process.
Join Chris and I as we dive deep into the world of AI. We chat about AI’s impact on creativity today and what the future holds for advertising and the arts.
Zack Seckler: So Chris, let’s start out talking about your history in advertising and what you’re doing now at Adobe.
Chris Duffey: First off, I think what you’re doing is a real service and compliment to the creative community. You’re not only a content creator, but you’re a real advocate for creativity by documenting the creation process.
I recently joined Adobe—I’ve been here about 7 months now and it’s been inspiring to be part of a company that fundamentally celebrates creativity and technology. Previously, I was on the creative agency side for 20-plus years, and rose through the ranks to Executive Creative Director.
You started out at Leo Burnett for Nintendo?
Yes, Leo Burnett, 35 West Wacker Street in the great city of Chicago. Starting out working on Nintendo at LB, was a kid’s dream. Then I made my way to Madison Avenue and have been here in NYC ever since.
So, “AI enhanced creativity”—that’s the term you used in your talk at Cannes Lions. Tell us what that is.
AI and creativity – music to my ears. Before diving into AI, it’s helpful to take a step back and talk a bit about the industry and cultural/market dynamics as a whole, and then show how AI can address some of those challenges. And there’s a lot of market dynamics and changes happening at the moment, but they all seem to bubble up to three universal sets of questions.
Number one is Data—there’s so much data, what do we do with it all? I heard a great fact the other day: humanity creates close to 250,000 Libraries of Congress per day. So now, it’s a matter of how can we best leverage and unlock it all and extract meaningful insights from it?
Number two is Content Velocity; the need for more content, faster. How can agencies and brands create more content? Yet not at the risk of compromising creativity or craft?
And number three is Innovation: there’s so much new technology, how do creatives, agencies and brands stay cutting edge with it all and create a constant culture of innovation?
Talk a little bit about where we are today with AI and AI enhanced creativity.
Essentially, there are three types of AI. There is Artificial Narrow Intelligence, Artificial General Intelligence and Artificial Super Intelligence. And to this day there is only instances of Narrow AI – and that is by all known accounts modern AI – which is very task focused.
Adobe Sensei, the company’s AI and machine learning framework, is the technology that powers intelligent features across all of Adobe’s products to improve the design and delivery of digital experiences. Sensei is harnessing literally trillions of content and data assets – everything from high-resolution images to customer clicks to solve digital experience challenges. And this where AI is very exciting and is going to help creatives get to solutions in a more profound manner, quicker. It’s under this notion that machine and human together are more powerful. We fundamentally believe that AI done right will amplify human creativity and intelligence – not replace it.
Give us a couple of specific examples, like in photography or video editing. I’ve seen that recently there was a whole movie trailer that was edited with AI…
AI and machine learning is great at sensing and predicting, specifically around pattern recognition. It’s very good at identifying a pattern whether it be on a pixel level or in written form or spoken narrative. AI can then identify actions to mimic those patterns and ultimately in the case of something like an AI creative assistant – or in the case of an AI assisted generated movie trailer, the system can help and be suggestive so the content creators get to solutions quicker – but with that said, for the most part AI can’t do highly creative and ideation outputs solely on its own.
Take Adobe’s “Make a Masterpiece” campaign, where we asked several of the world’s leading digital artists from the Behance community to recreate lost, stolen or destroyed art. Using Adobe Stock assets coupled with features powered by Adobe Sensei, these technologies helped to comprehend the composition and then identify images that would aid in recompositing those masterpieces. Plus, it assisted the human creators to place them. And that’s just one example of AI enhanced creativity.
Could we get to a point where a human could choose, say, 100 images and go, “Okay AI, you place them”?
I think we’re seeing glimpses of it in the automation of multiple renditions of core assets. For instance in templatizing variations of banner ads.
What do you think things are going to look like in five years? I know that’s an eternity in tech, but what’s on the forefront that you can share with us?
At the recent Adobe MAX creativity conference, we showcased a prototype of the Creative Graph, which was a pretty amazing glimpse of what content creation could look like in the very near future. Essentially, capturing and mapping the creative decision making process—and then at any point, the creator can go back to any decision and can alter that decision to see how that would have an effect on the ultimate output.
Is this in relation to video editing?
In this instance it was image creation and manipulation, but it could have a parallel effect on video creation as well. Specific to video, Adobe Character Animator was just released – it allows for the creation of a 2D character to come alive by mimicking your facial movements – so the characters act and react very realistically, interestingly the machine learning system was trained by watching hours and hours of tv. Even The Simpsons and The Late Show have put Character Animator into action!
What excites you most about the future?
That’s a good question. I think the notion of a CyberHuman Creativity Loop is really exciting. We were chatting earlier about the balance between craft and volume, and I think with an AI we’re going to be able to get to the sweet spot of addressing both.
That’s on the content creation standpoint. And from a distribution, measurement, and optimization standpoint, AI is having an impact on the overall consumer experience as well, where brand interactions and moments will be more contextual, more personalized, and ultimately more meaningful.
Imagine an in-market place campaign—those performance data points can be surfaced almost in real-time to the creatives. And it’s not just statistical data, it’s meaningful extracted insights that are both subjective, objective, and predictive, across all of the assets of that campaign. All with the goal to optimize content more insightfully and more efficiently in a continuous loop cycle.
So to use the example of banner ads, actually it’s funny as Jason M. Peterson CCO at Havas was just talking with us about how much he hates banner ads [laughs]—but to use that example: you could have three banner ads, and it would say, “Ad A is doing better than ad B, and the reason is probably because the BUY button is in this spot…” and then creatives could update that.
Regardless of medium or channel, that’s exactly it. And that example is just on an aesthetic level. Over time there will be a corpus of data on a more holistic view of an asset, from both how it’s performing as a message and medium; and why it’s performing well or not within the context of the overall customer journey and market landscape.
This is fascinating stuff, very business-oriented, which I enjoy talking about. But for somebody who’s an artist or a creator—who works in music, video, photography—tell me why they should be really excited about AI. What is it going to be able to do for them?
Without a doubt, it’s already getting the grunt work out of the way so creatives can do what they do best and be creative problem solvers.
Can you give me an example for photography?
Deep Cut Out for instance, is a Photoshop model that runs in 500ms with 92% to 96% accuracy on people, it can identify an object and silhouette it out. I remember back at Leo Burnett I would walk home some nights feeling like my eyes were bleeding from doing a lot of silhouetting. Being a photographer, I’m sure you know sometimes that’s a potentially tedious task. So imagine that in a 500th of a millisecond the system can identify the object, and now the creatives can go back to conceptualizing.
And what about video?
We’ve got a couple of projects going on where we’re doing object recognition with videos and images to auto-tag them for a number of different use cases. A basic example of how picture captioning assets can be helpful is for future reference and search-ability, based on very specific sets of criteria. Which by the way if this was done to this level of detail by hand would be extremely time consuming.
What about other technologies—augmented reality, virtual reality, 3-D motion graphics, stuff like that?
We just launched Adobe Dimension, which demystifies 3-D content to a great extent. A lot of 3-D software is very complex, and Dimension now allows you to take a 3-D asset and bring it into a composition and adjust accordingly. Dimension is a reflection of the need for very light, very fast and very user friendly application interfaces.
I know this isn’t really your field, but I’m curious for your opinion: what do you think is going to happen to the people who do the grueling work right now? What’s going to happen to all the people out there who make a living just silhouetting or whatever?
There’s always going to be the need for the human. In the marketing space AI is going to allow creatives to effectively address content velocity. With this need for more content, the human operator when assisted by AI can now focus on being more creative as well as creating it faster.
So that brings us to a question about creating content in general: there’s been a content explosion, as you were saying, and the quality has gotten better and better and people are becoming better at creating quality content. What is a future going to look like when it becomes incredibly easy to make something beautiful? If I can just cue something up with AI and quickly make engaging content, where do you think that is going to place professional artists?
I like to think of it as creatives are moving into this era of being Experience DJs. We’ve got a number of tremendous and various content creation technologies literally at our fingertips, and now we can start to creatively mix and sample the narrative and brand story in epic ways. The newly released Adobe XD is a great example of this.
From a creative standpoint, how is advertising going to change in the next five years? What about some of the newer technologies?
We’re very excited about the combination of Augmented Reality with AI, in the sense that these technologies are going to be able to enable customized, in-location contextual experiences on a very individual basis. Beyond personalized even. It’s predictive and suggestive—based on where you are, where you’ve been, and where you’d like the experience to go.
More broadly speaking brands are going to get so in touch with their consumer’s needs through AR and AI that it’s going to form a very deep, one-on-one relationship with their customers.
How can content creators and artists best utilize AI today? Any suggestions for staying ahead of the curve?
Just get into it. This is reminiscent of the early days of digital, similar to the mid-to-late Nineties. Some of the greatest winners were the practitioners who just got out there and started to experiment – now is the time to get in there with AI on a very practical hands-on level.
For example, if you look at Adobe’s Sensei, there are now more than 100 Sensei-powered capabilities in product or in our pipeline across Adobe Creative Cloud, Document Cloud and Experience Cloud. These features are helping our customers work smarter, faster, more efficiently and effectively – whether they’re designing an app, editing a photo or film, or launching a marketing campaign.
So I know there’s a lot of exciting stuff, but I know there are a lot of concerns about AI too. One of the big questions is, is there ever going to be an algorithm for creativity? What do you think?
It’s an interesting question. As we talked about AI is currently very good at pattern recognition, and with that in mind, the definition of creativity is that it’s essentially a marriage of unexpected correlations. So, there could be an AI assistant that helps spark ideas by delivering complex pattern recommendations of unexpected correlations.
Can you tell me what that might look like?
Adobe is already experimenting with AI as a creative and brand assistant. It could do a sentiment analysis of an asset and then indicate where within a predefined brand personality spectrum does that asset sit.
And, on an individual creator level, AI can also learn from a user’s behaviors, actions and preferences and over time can be suggestive and predictive based off of the color palettes you traditionally have used, images that fall within your brand guidelines, typography that you use quite often—suggestive and predictive algorithms are definitely going to be used in the near future for creativity.
Things have been changing a lot in the ad world recently. From your advertising experience, how do you think the agency of the future is going to change? Is there going to be any correlation with AI?
Without a doubt. I bumped into Keith Renihard, chairman of Omnicom’s DBB worldwide, at the airport a couple of weeks ago, and we were talking about how things have changed since I was there in the late 90’s. He said it couldn’t be a more exciting time to be in the advertising world, because we now have the ability to truly unlock insights on a greater scale. Advertising has always been based off of insights, whether it be marketplace insights or intuitive insights. We’re now finally able to leverage those insights and translate those into even more meaningful experiences.
For instance, imagine if you’re in the market for a new car. You’ll probably go online first and do some searching, and through this search and discovery process you narrow it down to a brand and model. Now, imagine over the course of the following days you are delivered social posts from a fan page, you are then offered to use the brand’s app to take a virtual test drive, next you jump to the brand website and are encouraged to customize your model and even vote for future features as well as see videos documenting the manufacturing process, later you come across a product placement in your favorite streamed show and a series of other digital and physical touch points. All of which reinforces the affinity you have for this specific automotive brand and model. And then imagine you go into the showroom and you are greeted via your mobile by an AI enable chatbot as well as a human sales person to answer any outstanding questions. In totality this sum of all these brand interactions that have been specifically tailored to you have created an extraordinary personalized experience. You then go home, and you get a thank you txt with a link to an owners community that showcases their uploaded images and videos. Collectively, this individualized experience across all these channels have amplified the overall brand value. Through this marriage of art of content and science of data the brand delivered deeply understanding moments and that’s the power of AI enhanced creativity.
Interview by Zack Seckler
Edited by Francis Carr Jr.