Celtra had an incredible time celebrating cutting-edge technology at RISE 2019! During the event, we held a workshop, “The Possibilities of Creative AI”, which included a panel discussion led by Raushida Vasaiwala, our General Manager APAC. The industry-leading participants were asked a series of fascinating questions about the future of Creative AI and its role in marketing. In case you missed it, or want a recap, we laid out the standout QAs.
You can also watch the full discussion here.
“AI comes from humans, and humans are inevitably affected by their environment, so bias is always there. But if we can teach ourselves and machines to be more considerate about different ideas, we can make a lot of progress.”
–Anthony Baker, Executive Director, Technology, R/GA
Raushida Vasaiwala: Do you think AI can be taught to be creative? Do you think it can operate without human guidance on the basis of pixel arrangements or color palettes?
Anthony Baker: It’s an interesting challenge. I think we are talking about several important things. Data from creative experience, where we’re talking about using creative in a more emotional way. Then there’s sensibilities, understanding something more emotional than demographics. (…) At Cannes, we used open AI gbt-2, an algorithm-like breakthrough in text creation. You tell a phrase and it will write the whole story for you. Once it’s trained, you can prompt it with a sentence or question and it will write a case study for you. Sometimes it gets it really wrong. It doesn’t understand sensibilities. (…) AI comes from humans, and humans are inevitably affected by their environment, so bias is always there. But if we can teach ourselves and machines to be more considerate about different ideas, we can make a lot of progress. But we’re not there. We can teach AI to follow the masters and they can imitate that type of style.
Raushida Vasaiwala: So Anthony, who will solve Creative AI first? Will it be platform companies like Alibaba or Google? Or creative companies like Adobe or Celtra?
Anthony Baker: Hopefully it will be me. If not, it’s hard to say. There are different verticals, angles, perspectives. I would definitely say Adobe’s putting a lot of effort and investment into generating creative as content and art, which is fantastic. On the other side, whether you consider it Creative AI, Alibaba is using their AI and machine learning technologies not only to generate creative, but to test that prototype, send it out, create projections for companies to know how much they’ll make from a product that doesn’t even exist. But at the same time, companies like Google are doing a lot of experiments with AI that impacts how other companies can use it for breakthroughs. (…)
Qosmo in Japan is a DJ AI. It’s not meant to replace DJs but to perform with the DJs. It integrates based on audience response for a better performance. So these kinds of things are very interesting. Then you have you guys [Celtra], using these technologies to create something effective. Something that is actually there, reaching people, and elevating creative intelligence in the industry. (…)
“I think the more Creative AI power and the more we use it right, the better for economies, the better for the brands, and the better for the whole ecosystem.”
–Narayan Murthy Ivaturi, Global COO, Freakout Holdings
Raushida Vasaiwala: When talking about all the possibilities, are you afraid of Creative AI?
Narayan Murthy Ivaturi: No, I don’t think so. What you’re describing is cognitive thinking. I don’t think machines are anywhere close to this. You can train a machine to do ABC, XYZ things. But you can’t tell them when they should do it. That’s cognition. The human brain is made up of thousands of neurons which create the complexity of the human brain. The way it is now, I don’t think it’s possible to recreate 1/10,000 a proportion of what they [human brains] can do yet. Yet…yet, yet. That’s all fiction now. All in movies right now. It’s fun to think about the point of singularity but I don’t think we’ll get there in our generation, at least. (…) It doesn’t scare me, it’s a tool. What scares me is that people are not trained in how to use the tool and that can create bad experiences for the client. The scare is not the machines, it’s the people using it in a bad way, creating a bad impact that will create a negative overall impression of this technology which has been happening for some time now.
Raushida Vasaiwala: So Narayan, when we speak about creative and media, how can they be converged? How can it help achieve economies of scale?
Narayan Murthy Ivaturi: It’s happening right now as we speak. More and more media is bought programmatically and getting mixed up with whatever data triggers we have — location weather — qualifying the user in that moment. Economies of scale are more concerned with advertising effectiveness: what is the consumer offtake? What is the bounce rate? What is the brand lift? This is where Creative AI is more essential than anywhere in the market. Creative AI is making sure the consumer is getting the right message at the right time, in the right manner, with the right color, pixels — whatever we choose. The engine is getting up to deliver at the right time which is the essence of the advertising we all learned. (…) With the right data, intent, and time, conversion will be much higher. I think the more Creative AI power and the more we use it right, the better for economies, the better for the brands, and the better for the whole ecosystem.
“It would be a mistake not to evaluate AI based on what it can do today. There’s still the human element but that doesn’t necessarily mean that will be the case in the future.”
–Philip Tabet, Managing Director APAC, Mobkoi
Raushida Vasaiwala: From a future standpoint outlook, what are you most excited about?
Philip Tabet: Dynamic creatives are the future when it comes down to real-time optimization of creatives. A great case study about this with a good intersection about focusing on a new idea comes from a collaboration between IBM and the movie “Morgan”. (…) The project was can an AI create a trailer using algorithms? AI had to be taught to interpret and use data to identify parts of the movie with emotion. Then, use this information to put it [the trailer] together. The AI algorithm used here is something called supervised learning, when a human has to teach a machine how to interpret deeper emotions. It has practical uses in industries like retail, medicine, and finance. (…) It would be a mistake not to evaluate AI based on what it can do today. There’s still the human element but that doesn’t necessarily mean that will be the case in the future.
Raushida Vasaiwala: Phillip, how would you tie the consumer experience to Creative AI? Which brand is doing it really well today?
Philip Tabet: I think it’s more industry, not brand. There needs to be a new system in place to help us adapt and learn. There needs to be a transition soon, from data-driven experience to something that is more human. (…) The cosmetic industry is buying a lot of technology and looking to connect more deeply and personally. SHISEIDO is doing great. Our client, Estée Lauder too. We work a lot with them to implement facial technologies with customization as consumers interact with ads. Another example would be MAC. They’re going another step with an experiential store in Shanghai. They’re consolidating all the data points which are managed through a WeChat platform. They understand everything about you and curate an experience as you walk through the store.
“The tipping point comes when emotions can be transferred into data, and data can be transferred into emotions. The reality is I think we’re close to that tipping point with things like facial and voice recognition.”
–John Kim, Client Services Director APAC, Red Fuse
Raushida Vasaiwala: John, we spoke about dynamic creatives. How does AI help with consumer experiences specifically?
John Kim: When it comes to AI, we need to go beyond Creative AI. AI can be used for so many applications. Always put consumers at the heart. The power of AI will be to anticipate needs before the people even know they need it. Like on the Starbucks app, you use it every morning. Your coffee is ready on time when ordered through the app. Imagine every morning this happens. But one morning you have a meeting. Imagine the app uses your calendar to know you have a meeting, then it pings you that it will be ready early. This type of prediction will be what’s really important for the consumer experience when it comes to Creative AI.
Raushida Vasaiwala: Does AI scare you?
John Kim: I have a slightly different point of view. I think we’re all on the same page in that we’re not there. With all great tech there’s always a tipping point. And for machines to act like humans, machines need the ability to feel like humans. Humans are extremely complex — think about hand gestures, words you choose to use, how you explain yourself. The tipping point comes when emotions can be transferred into data and data can be transferred into emotions. The reality is I think we’re close to that tipping point with things like facial and voice recognition. I think we’re dancing on that tipping point on a day-to-day basis. Yes, I think we’re close to that tipping point and I question how will it happen and what impact will that have on advertising. My greatest concern is then creative will be commodified. Creative would lose its edge.