BodyBackground

Projects

Allison Crank – THE BEST PART OF POSSIBILITIES: ADVERTISING MADE BY A BOT

ONLINE ADVERTISING AND AI DESIGN

 

PROJECT

Advertising seeks to transform products from virtual possibilities into physical reality. In that sense, the goal of advertising is to produce consumers rather than products. Marketing campaigns strategise to make certain goods real and competing products virtual, constantly changing the design landscape. But in recent years, the quantity and variety of products, vendors, and consumers on Alibaba has led to a specific demand—for engaging, intuitive, and user-specific advertising generated through artificial intelligence. Scientists at Alibaba’s research centre are developing AI tools, including neural networks, to achieve that goal. In that sense, Alibaba is becoming both designer and producer—a consumption machine, blurring the boundaries between virtual and real.

But what lies behind these artificial campaigns, and how does this virtual tool create demand in a physical world? The Best Part of Possibilities is a self-named neural network bot that has been trained on thousands of American advertising slogans from the past 100 years. These amusing, nonsensical, and cliché phrases, assembled into new phrases using machine learning, reveal the uncanny yet irrational pattern and syntax of advertising which we have assimilated in the era of broadcast and interactive media.

 

INTERVIEW

What did you learn about Alibaba’s approach to AI and advertising in your research and design using a neural network?

Alibaba’s researchers are focused on developing new methods in artificial intelligence and publishing their breakthroughs to add knowledge to the field. The fact that their work is eventually applied to advertising is much less important: their job is not to speculate about the ethics of their impact on consumers but to drive technological innovation. By working on an advertising neural network, gathering inputs and creating the code, you also start to see it like a bot—more in terms of patterns and performance. It takes time to train a neural network: you are constantly testing, adjusting, and fine-tuning the framework. Our first outputs were almost completely incomprehensible and couldn't be used as texts, but the longer it ran, the better the outputs became.

Is there a relationship between the design of advertising and the design of products?

Most advertising is never really about the product, but rather about you, the consumer, and a specific possibility. Advertising tries to forge an artificial connection between you and the product. But as product design develops more towards unique objects or hyper-customised mass-market objects, it puts pressure on advertising to change with it. It's no longer "buy this product to make your life perfect" but rather "make this product be perfect for you."

The design of your video is a mix of Las Vegas and Times Square; how do you see the future of the environment in which advertising will be received?

AI advertising is trained to target you specifically, calling your attention to get you to click and buy. But personalised advertising can still be annoying, like spam emails based on your browsing history or text messages for discounts when you physically walk by a store. It can becoming overwhelming and Big-Brother-like, trying to sneak in and infiltrate your daily life. I hope the future environment of advertising can be a bit more honest. l like the visual language of Las Vegas and Times Square because even if the buildings are also trying to call your attention to enter and buy, they fully embrace and celebrate their existence as advertisements. They are immediate, in-your-face, but transparent to the viewer about their commercial nature. That approach to the visual environment will only become more significant as AI advertising blurs the line between the product imaginary and your reality.

Project Images

Copyright Design Academy Eindhoven. Photographs by Nicole Marnati.