Creativity in the cloud: Is AI-fueled optimism justified?

Artificial intelligence’s ability to transform creative working practices has been thrust into the spotlight of late. When CHI & Partners chief executive and partner Sarah Golding launched her IPA presidency on a tech-focused agenda in April, she revealed plans to roll out IPAi—a platform aimed at encouraging experimentation among member agencies and a training programme to demystify what she called the "history and hype around AI".

Then, in June, during the Cannes International Festival of Creativity, the industry was rocked by the news that Publicis Groupe is bowing out of the event and cutting its entire marketing budget in order to invest in a new artificial intelligence platform called Marcel.

According to Joe Stanhope, vice-president and principal analyst at Forrester, the "hype" and "consternation" surrounding the Marcel announcement is disappointing.

Such a reaction, he argues, often ends up dampening enthusiasm for a transformative technology by sending expectations skyrocketing and setting the stage for disappointment.

Nonetheless, Stanhope believes that AI is a game changer for the marketing and advertising industry, and will herald a "renaissance in creativity".

"It’s important for brands to become knowledgeable about what AI is," he says. "It will become increasingly important and valuable, and they will want to be ready for that. Anyone in an agency who complains [about the time and resources that AI can save] isn’t thinking about their client."

Algorithm for creativity

Stanhope is not alone in his enthusiasm for and optimism about AI. Experts are happy to reel off a long list of the reasons that AI will prove such a positive force for marketing.

They say, for example, that it will bring many brains to briefs, from around the globe, in an instant. Moreover, these will be the best brains, based on past experience. It will support the diversity agenda. It will enhance creativity by providing a robust framework for it. It will transform disparate groups into true, collaborative networks. It will combat "data overwhelm". It will make agencies far more efficient and streamline the costing of jobs. It will save time and human resources.

Practitioners working with the technology are buzzing not only about its potential, but also the benefits it already offers marketing. Take IBM. It is famous for creating the world’s first cognitive computing system, Watson, known widely for its use in oncology, as well as writing pop ballads, mimicking the artistic styles of great painters and taking on the role of film director.

Lisa Gilbert, chief marketing officer at IBM UK and Ireland, says that the company has also made "great progress" by employing AI "as inspiration in the creative process". A prime example is the use of Watson in IBM’s programmatic media buying to improve the relevancy of its ads to its target markets—the system continuously learns about optimisation.

The average worker spends 28% of the working week managing email and nearly 20% looking for information internally

Gilbert’s team is running several other AI pilot projects too. For instance, IBM has plugged its "Cognitive Bid Optimiser" technology into its demand-side platform for its paid-media buys, to optimise bidding on ad groups. It has also used AI to optimise on browser type, leading to a 33% reduction in cost per landing in the initial phase of testing. In addition, the IBM marketing team is now experimenting with its own cognitive bot to enhance the customer experience online.

"With the power of cognitive insight on the marketing team, personalisation becomes sharper and more refined, improving the quality of every customer interaction," Gilbert says. "Cognitive computing is a new level of collaboration between man and machine, and will only augment human intelligence, not replace it. This is man working with machine to help humans do their jobs better."

The 100-year transformation

Fully automating the laborious data donkey work frees up marketers to focus more strategically and use their time more efficiently and effectively. Johann Butting, head of EMEA at team-collaboration service Slack, predicts that this is just the beginning of a "100-year transformation of work", with AI at its centre.

He cites AI’s ability to facilitate stronger collaboration by automating and simplifying "rote tasks" that are key to the process but "essentially a distraction". He also points to McKinsey research that shows that the average worker spends 28% of the working week managing email and nearly 20% looking for information internally.

"We’re just starting to understand and capitalise on what could be possible in a workplace enhanced by mobility, bots and AI," Butting says. "If we can help unlock even a small increase in productivity, or automate tasks that are just a drain on mental bandwidth, then we will make a major impact."

He singles out car-sharing company Zipcar as a client example. It uses Slack to generate ideas and pool individuals from different departments to boost marketing innovation.

Elizabeth Closmore, global head of product evangelism at social-media management › platform Sprinklr, is similarly effusive about AI’s ability to take away the heavy lifting work from marketers. She goes as far as to describe it as "the most exciting thing to ever happen to marketing, because it makes the creative work you put into any piece of content more relevant", adding that "you can evaluate the potential of every marketing campaign in advance".

Far from making creative professionals redundant, she argues, the AI revolution will serve only to increase demand for creativity and content to feed the machines. It is a view echoed by many experts, including Stanhope. "It’s going to put a premium on creative and content. We’re going to need more than ever," he says.

AI is awesome

So far, so good. At this point in proceedings, the theme tune to the AI-powered workplace story sounds as if it should be a variation on The Lego Movie song Everything is Awesome.

Stanhope articulates this: "AI can work out what works and what doesn’t way before a human could, which is really cool, because that’s a hard, thankless task. I would love my computer to work out while I’m sleeping that ‘this promotion worked really well, this one didn’t’. That would be awesome."

And it is. Or, at least, it could be. But—and here’s the potential fly in the ointment—these systems will flourish only if man works in harmony with machine. It’s not as simple as spending millions on an AI platform, before presenting it with a "ta-da!" to the workforce and expecting it to take off. The part played by human engagement is at least as important as that played by tech development. After all, what is the point of having the best tech in the world if no-one wants to use it?

This may sound obvious, as PHD’s worldwide strategy and planning director Mark Holden attests, but it’s not. He should know; he’s the mastermind behind PHD’s gamified collaboration platform Source, which he calls an "incredible undertaking for us". While he applauds Publicis’ aspirations for Marcel, he cautions against setting expectations too high, especially when AI technology is in its infancy.

There isn’t a HAL 9000 you can go to and say, ‘Hey, can you put together this new campaign for me?’ It doesn’t exist and it’s going to be quite a long time, if ever, until it does.

"People aren’t always that interested in helping people from other parts of the world on briefs that they don’t themselves work on," he says. "It has taken us a number of years to get Source right. Getting things functioning in a way that [means] people actually use it is difficult. You can spend a lot of money on a piece of software, but you’ve got to bring your people in and change behaviours."

Holden views Source as a work in progress, with great potential still to be fulfilled, not least the likelihood that AI functionality will be added to it. His current focus is predominantly on getting more people within PHD, as well as clients, to use the platform. While Source has won awards and had a positive impact in the workplace, only 1,500 employees out of a possible 4,500 use it each week.

The human obstacle

AI platforms—for now, and the foreseeable future—are only as good as the humans operating them. If said humans are disillusioned, disenfranchised and disengaged, a platform aimed at collaboration won’t work. The problem with AI—such a loaded word in marketing—is that the human emotions it provokes risk getting in the way of machines making a difference.

Stanhope has undertaken research on the myths surrounding AI. He discovered that we bring preconceived notions garnered from science or popular culture to it, and these automatically pop into our heads when we think about it.

"You can’t help it," he says. "I always think of 2001: A Space Odyssey [when the computer, HAL 9000, takes over from the humans with fatal consequences]. But I know that, in reality, AI is completely dependent on the information we give it and how we train it to solve problems. AI doesn’t do things on its own. There isn’t a HAL 9000 you can go to and say, ‘Hey, can you put together this new campaign for me?’ It doesn’t exist and it’s going to be quite a long time, if ever, until it does."

Nevertheless, the subconscious is the brain’s own powerful machine. Add to this the fear that "the robots are coming to take our jobs!" and it is perhaps no surprise that AI activates the natural human negativity bias in the minds of many marketers.

So, when introducing a system such as Marcel, the humans for whom the success of an AI collaborative platform is intended must be carefully managed. If it is implemented with care, openness and patience, they are more likely to be receptive to—and come to embrace—the tech, not resent it.

Fuelling passions across platforms

Linda Hill, a professor at Harvard Business School who specialises in innovation and talent management and is co-author of Collective Genius: the Art and Practice of Leading Innovation, believes AI platforms can be hugely powerful tools for collaboration and creativity. In her experience, these platforms work best when employees are given the choice to work on topics that are of genuine interest to them, they are well supported and allow talent to be seen as "global" rather than "local" across the network.

Additionally, she recommends identifying "a way to make these platforms be a place where employees can also intermingle non-work activities—more like social networks".

However, most important of all, she advises that this collaborative, AI-fuelled future requires a new type of creative leadership, one that operates from the bottom up.

"Leaders need to know how to amplify differences, work with passionate people and deal with creative abrasion, and the conflict that will come from that," she says. "It requires a completely different leadership mindset to take advantage of these types of platforms. It’s not about being visionary. Hierarchical cultures actually make it harder to take advantage of the benefits."

Despite the initial obfuscation and secrecy, Publicis Groupe is now opening up Marcel. By testing possible uses for the platform through its M Labs incubator, chief strategy officer Carla Serrano says "many people" will have touched the project by June next year—from account management to tech. She also hopes to be open about the failures. "We need to be transparent about what works and what doesn’t," she adds.

In the meantime, founder of collaboration specialist Let’s Go and former ?Whatif! consultant Richard Watkins leaves those at Publicis with some advice.

"You can’t just data-mine ideas. Therefore, success will depend on the quality of conversation that Marcel can create, not just the individuals it can connect. Fine, it connects brains. But how does it connect humans? If they’ve just got a couple of clever people trying to figure this out, they will inevitably build a broken system," he says. "I hope they build it collaboratively, because it has so much potential. If they don’t embody what they’re trying to build in how they are trying to build it, they will fail."