Artificial intelligence will kill design’s ‘Hippos’

Anyone who has worked on design knows how human the process can be, but artificial intelligence can help bring some rigour to what can be seen as a subjective area.

designAt an Econsultancy breakfast briefing at the start of this year, I was struck by what panellist Gregor Young said about artificial intelligence (AI). Young, who is currently leading the transformation of Channel 4’s digital marketing strategy and capabilities, emphasised how widespread he believes AI will become and spoke of how frequently he encountered processes or problems where he wished he could apply some form of AI.

Young’s enthusiasm made me revisit an article I wrote last year, in which I asked, ‘Why are marketers kidding themselves that AI is about more than sales?’.

At the time I was simply making the case for a very pragmatic approach to machine learning, given that many of its current uses – recommendations, product categorisation, personalisation, copywriting – are all about optimising for conversions.

But what I didn’t cover was a swathe of algorithms being used in digital and graphic design. The use case that most effectively drives home just how quickly marketing may be disrupted by AI-powered design is demonstrated by Alibaba’s LuBan AI platform.

Few companies can demonstrate effectiveness at scale quite like Alibaba, which reportedly used LuBan during the 2017 Singles’ Day shopping festival to design 400 million banners to advertise a range of products.

Anyone who has ever worked on a website design and received unsolicited or unfathomable feedback from a senior manager knows just how human design can be.

In a blog post about the platform, user experience designer Ke Xu estimates this output to be equivalent to the work of a human graphic designer working non-stop for 150 years.

What starts with banners can be extrapolated to all sorts of UX design. Though the process isn’t entirely automated – designers provide feedback to the algorithm as well as evaluating final products – the implications are clear: Ke Xu’s blog post is provocatively titled ‘AI visual design is already here — and it won’t hesitate to take over your petty design job’. Ouch.

Is the narrative that data and creative are opposing forces true or indeed helpful?

Well, design is undoubtedly a very human process. As Yury Vetrov, design director at Raiffeisen Bank, writes in a 2017 article for Smashing Magazine: “Designers make a lot of big and small decisions; many of them are hardly described by clear processes. Moreover, incoming requirements are not 100% clear and consistent, so designers help product managers solve these collisions — making for a better product. It’s much more than about choosing a suitable template and filling it with content.”

Anyone who has ever worked on a website design and received unsolicited or unfathomable feedback from a senior manager knows just how human design can be. We all think we can do it, because we feel there is a certain amount of instinct involved. It’s why the tweet below is the funniest of the year – it shows a designer’s price list increasing in proportion to the level of stakeholder involvement, with a lovely punchline at the end.

It turns out that proving the value of a particular design to management is one major selling point of some AI-driven design software.

Charles Blake-Thomas, CEO at Eyequant, a company that provides visual attention analysis software and can predict which UX designs are most likely to convert, is frank about this dynamic: “If you look close enough you’ll uncover that the brands that have the most cluttered websites and worst user experience all have one thing in common. And it’s not an unskilled designer or inadequate software or methodologies.

“In reality, all of these websites have a Hippo calling the shots. A Hippo (highest paid person’s opinion) can have disastrous consequences for UX as design decisions are made without user data and are based on nothing else but opinion which, unbeknown to the Hippo, includes cultural preferences and cognitive biases.”

It feels like AI-powered design is not evenly distributed at the moment, but that its incursion into marketing is inevitable. However, though the writing might be on the wall for the repetitive work of versioning banners and testing them, designers will still have jobs; they will simply involve greater collaboration with algorithms.

Microsoft interaction designer Jasmine Oh writes in a blog post: “While AI will replace designers, it will replace the designers of today, not the designers of tomorrow.

“AI will become a design partner and tool that designers can use to meet ever-evolving workplace demands. And when nurturing any relationship, let’s learn what our partner can and can’t offer.”

That brings us neatly to the pressing issue for marketers: they need to learn about this stuff. Not just the technical side but the ethical side, too. Netflix’s personalisation hasn’t been without controversy, for example – some believe the streaming giant has provided misleading representations of titles to attract viewers based on ethnicity.

Smart martech companies in this space know that a lack of understanding is holding marketing departments back. Phrasee committed to educate 500 marketers in 2020 about the AI opportunity.

As CEO and co-founder Parry Malm puts it: “AI is a buzzword; it is difficult to cut through the hyperbole and understand what’s actual AI and what’s BS. And it’s hard to know the tech that will make a real difference and where tech exists for tech’s sake.”

All of this adds up to a lot of excitement for designers and marketers, and hopefully the beginning of the end for Hippos.



    Leave a comment