CMO.com recently published this article by our Innovation Director, Tom Ollerton, looking at the intersection between creativity and AI in advertising - and where it will lead us in the future. They've been kind enough to let us reproduce it below.
On the 6th of June, the New Scientist tweeted the following message to its readers:
What Musk didn’t know was that the New Scientist tweet was written by an AI. That AI was the creation of Echobox, a tool that writes and posts tweets on behalf of publishers using public and private data to predict virality with deep learning and neural networks.
If the Cristiano Ronaldo of AI can be fooled by an algorithm, then marketers need to ask themselves: What can AI do for my brand?
Many in the industry doubt that the AI creative singularity is coming anytime soon. According to the software company Sysomos, 63% of marketers feel confident creative jobs will prove to be resistant to the threat of automation. But a closer look at some exciting AI startups such as Echobox will show us that machines can be creative enough to fool us that content came from a human. So how soon can business leaders use this tech to sell more?
Man Meets Machine
The tension between the potential of AI to replace some creative processes and the industry’s aversion to change is palpable as we progress towards a human and machine-driven industry.
Benjamin Lickfett, head of technology and innovation from the drinks giant Diageo, has been exploring the use of AI and voice based search and spoke on the subject at recent innovation event, I’ll be Back. The brand’s main foray into algorithmic marketing is “conversational commerce,” which he believes is a big opportunity for the brand that it is exploring for its Scotch portfolio. The nuances of the category are difficult to understand, and chatbots have helped, but they are a “glorified decision tree” as opposed to true AI. But Diageo didn’t stop there. In Cannes this year, it had a voice-powered bar, using Alexa, where you could order your drinks from the table.
Lickfett also gives a nod to the consultancy 10x, which had created a beer brewed by AI. It releases new variants regularly and gathers reactions to the beer on Facebook Messenger to inform the next batch. Lickfett concluded that this could be what the future of product development might look like.
The theme of AI and alcohol is also referenced by Rob Saunders, researcher in computational creativity, who is one of the most influential academics in the computational creativity space. He has taught his students how to use AI robotics to create machines that can make cocktails. However, his latest work is focused on using artificial intelligence to create “disposable games.” The aim he claims is to use this technology to help brands and consumers create games that are played only for a few minutes, before being forgotten about in the same way as an ephemeral photo or video content is produced.
Into The Written Word
The opportunities for marketers to get machines to produce their content extend far beyond gaming and specifically in the creation of the written word. Maria Flores Portillo, UK general manager at language generation company Persado, believes that AI today was not that sophisticated, it was “just a faster way of doing an analysis of data.” The main problem with AI is that it’s still limited. Currently, AI can be great at chess or driving a car, but it has no general intelligence beyond its skill. It, certainly, isn’t capable of creative intelligence—but it can copy humans.
Persado is harnessing AI’s ability to copy human patterns, and using this to create machine-written copy. The main advantage here is that machines can automate the typically menial, low interest “sales” copy—but they can also try multiple variants of that copy by tapping into core human emotions. Flores Portillo’s view is that AI should be able to help creative, and that marketers should use AI to challenge their beliefs about content.
If brands want to get involved in AI, they need two things—the right data and algorithm. The good news is that there is data aplenty. Antoine Amann, CEO and founder at Echobox, told the “I’ll Be Back” audience that, from the dawn of civilisation until 2003, only five exabytes of data had been created—now we produce five exabytes a day. Data is there to be taken, but, as Antoine tells, us “data is the new oil, don’t drown in it”—this is where the algorithm comes in.
We Have Been Wrong Before
The importance of automation in advertising is obvious to Echobox, who has staked its claim to the prize but also conceded that we had often been wrong about the future. We’re looking to wind power over nuclear, EasyJet not Concorde, and we still don’t have flying cars.
How should marketers think about AI and ads? Alex Hobhouse, strategy and innovation director at Saatchi & Saatchi, believes that the likely future of AI and advertising is that AI will take care of the grunt work, giving us more time to think in a way that the robots can’t. His excellent analogy of this situation is that the hugely successful Ford Focus is built primarily by machines, but, by contrast, Ferraris rely on human specialists working with machines.
While Musk may not be able to spot an AI-written tweet from the New Scientist, there’s still a long, but interesting, journey ahead of the ad industry as it tries to work out the robot’s role amongst us, humans.