Artificial Intelligence is past the point of being a buzzword. It is embedded in all industries, with many different applications, from streamlining complicated processes to enabling customers to try clothes on before they buy. Despite its impressive executions, there are still questions about the extent to which AI can be deployed in the creative industries, and what impact it will have on the workforce and audiences.
A panel of experts from the publishing, retail, tech, and marketing worlds sat down to answer those questions, for a session with The Drum at The Drum Arms in Cannes Lions, to determine whether AI can enhance the creative process for marketers. The panel came up with three use cases for AI – creation, iteration, and delivery – which enable it to contribute to the creative process, though not completely usurp the place of the human.
The panel noted that where AI shines is in sifting through data to spot patterns that would have been invisible, or seemed counterintuitive, to humanity. They cited the example of Netflix’s data-based commissioning, which has led to the creation of hit shows like House of Cards, as evidence that the inception of a creative process can be helped by AI. Aaron Frank, SVP, marketing, insights and strategy at Branded Entertainment Network, explained: “I think what Netflix is doing is interesting on a number of levels; they switched over to their algorithm on recommendation last year from a human based algorithm to a completely computer-based algorithm that's now making recommendations.
"They're also using all that data to see ‘hey, people love romcoms and are watching rom coms on our service. Nobody's made a new romcom in five years. Why don't we do that?’
He noted that the granularity of data Netflix has on its users allows them to recommend to content creators that they trim the length of a cold open, or the ideal length of scenes, in order to best appeal to that genre’s audience. It is, as the panel pointed out, AI being used to influence the creative process before any consumer has interacted with it.
AJ Frucci, vice president of concert and programmatic for Vox Media argues that AI enables marketers to predict where best their chips might be placed even before they start the creative process, in addition to then allowing them to fine tune their efforts: “At Vox Media, we use AI to improve performance based on real time data that can predict and optimize our content and creative formats via two tools: the first, Creative Intelligence, is used to test and optimize multiple versions of a creative format in real time against different audience segments and our editorial networks. The second, Conversational Intelligence, crawls the content of a page to detect key words, context, and emotion to improve brand safety and ad targeting, and soon this tool will allow us to match ad creative to content using AI.”
However, despite that immense power, the panel was keen to point out that AI on its own can at best remix existing materials, rather than create something from whole cloth. Henry Arlander, head of product and innovation at Pereira O’Dell used the example of AI that can create a ‘new’ Beatles song through analysis of their material: “You feed in the existing Beatles catalog and it can detect the patterns, how their songs are structurally created. And they can spit out a new version, but it's always going to be a Beatle song. It's never going to create like a human would, to take the next step to something.”
Eleanor Mills, the editorial director for the Sunday Times, agreed, relating the time the Sunday Times magazine’s art critic was sent to be sketched by an AI designed solely for art, and found it could not create anything that “any A-Level student” could not have also done.
While AI might not (yet) be able to create wholly new creative products, it is being deployed very differently for many marketing efforts. Thanks to huge data sets, new developments in facial tracking technology, and many other developments, AI can now be used to provide on the fly feedback to a piece of creative or marketing, massively streamlining a process that used to take weeks.
Mikhel Jäätma, the CEO and co-founder of Realeyes, developer of a tool that tracks a user’s facial responses in real time, explained that this enables the brand or publisher to iterate on an existing campaign, to ensure it is landing as effectively as possible:
“Essentially, Realeyes GO is a tool to quickly see if your video is good or not based on how much emotional engagement it generates and how much attention it attracts. It’s part of the creative team.
“You add your content to the self-serve dashboard, go have lunch… and when you get back, you’ll see what the emotion and attention scores are. You can actually see what connects with people.”
The great strength of AI, particularly for publishers, is in ensuring that the ads are being placed in the most effective place for the right audiences to see them. Debarshi Pandit, head of multicultural business for Sky Media noted that the great benefit of AI for advertising is spotting where particular audience segments feel underserved, and in catering specifically to them on a flexible basis: “77% of British Asians feel that mainstream advertising has got low benefits for them. And that's a staggering number when you think that there's almost 4 billion people who are not connecting with any other brands ads.” He cited the recent India vs. Pakistan test match as an example that delivered significant uplift to Sky in terms of new advertisers as a result of AI enabling advertisers to reach those audiences.
Benji Lord, the executive director for global omnichannel marketing, CRM and loyalty for Nars Cosmetics questioned whether much of what we term ‘AI’ should better be termed big data or some other name. He added that NARS is using AI to enhance the customer experience, by allowing them to ‘try on’ make up in a way that takes things like their skin tone into account. Despite that specific example, he pointed out that across different industries AI is more of a booster for existing technologies than an end in itself:
“It's tough today; everything and nothing can be AI really, so the only real pattern beyond machine learning that I'm seeing across industries is that AI is at the heart of a lot of technology fusion. So it's powering a lot of new other technologies whether it's social, whether it's augmented reality, you name it… it's interesting to see how AI is putting a lot of other technologies on steroids.”
In that sense, the development of AI tools to aid marketing is very similar to its deployment across other industries: it allows the creatives to focus on the business of creation rather than fiddling with the minutiae of delivery. Just as AI is enabling journalists, for instance, to focus on long-form articles while AI takes care of the rote business stories, for instance, AI is enabling marketers to be more creative, iterate faster, and deliver more targeted messages than ever before. Arlander explained: “I'm excited about that. And I know our creatives are too, because that will free them up to try to solve more complex creative challenges, which is really the value we provide for our clients.”
The danger is in thinking that AI on its own can fully replicate the functions of human marketers. It can’t – yet.
For more information on Realeyes GO, follow this link.