A Blog by Jonathan Low

 

Nov 29, 2018

The Problem With Invisible AI Branding

If companies wish to monetize the value added from artificial intelligence, they are going to have to brand it rather than hide it as a hidden, embedded element as they are doing now. JL


Jason Brush reports in Fast Company:

Consumers have few ways to understand when and how AI is being used, and to judge for themselves if they see it as a benefit or not. It’s not a recognizable element of a brand. If AI is to become identifiable and understandable by consumers, its value must be articulated. And for that to happen, designers of AI-driven experiences must make the invisible visible; they have to give AI a good brand identity.
Artificial intelligence is the most ubiquitous innovation you never see. It quietly powers automatic translation and closed captioning, automated media manipulation, search results, social media filtering, medical diagnosis, shipping logistics, and targeted advertising. There is likely no aspect of human industry and society that AI will not eventually touch–for better or for worse. Consumers, however, have few ways to understand when and how AI is being used, and to judge for themselves if they see it as a benefit or not. It’s simply not a recognizable element of a brand.
If AI is to become a meaningful facet of society, identifiable and understandable by consumers, its value must be articulated. And for that to happen, designers of AI-driven experiences must make the invisible visible; they have to give AI a good, old-fashioned brand identity.

The case for branding AI

A skeptic might wonder why AI needs branding in the first place. If it’s meant to silently toil away in the background of our lives, why does it need to announce itself? Why give consumers yet another thing to think about?
But brand-building has always been key to making impersonal industrial processes, technologies, and corporate organizations relevant and relatable to people–especially in cases where there’s little functional difference between one corporation’s products and services and its competition’s. Brands are the human face of an industry and the primary mechanism businesses use to make unique, differentiated promises to their customers.
Every successful brand articulates itself to people both through familiar visual branding assets–a logo, typeface, color palette, and so on–but also by crafting distinct experiences that make products and services tangible. The cushioning of shoes, the sound of a car engine, and the closing of its door, the color of a pill’s coating, the login mechanisms of a phone, the hidden chipset in a computer, the coating and insulation of a jacket, and the alert sounds of a digital device can all be meaningful (and trademark-able) elements of a brand experience.
Invisible by design
The key obstacle that AI faces, from a branding perspective, is that it has been engineered to be invisible. AI is often deployed as a way to eliminate friction and to reduce people’s awareness of technology. Unlike other familiar brand elements–color, typography, logos, texture, sound, tone of voice, photography style–AI is often seen as being most successful when it’s completely invisible.
Consider the role that AI plays in creating shopping or content recommendations within products and services today as opposed to, for instance, a sommelier at a Michelin-starred restaurant: a sommelier’s job is to converse, inspire, and educate; diners’ experience with their sommelier’s unique human intelligence is everything artificial intelligence isn’t: It’s highly visible, opinionated, and slows down a process rather than speeds it up. In contrast, Netflix’s celebrated, AI-based recommendation engine–which Netflix has estimated being worth $1 billion to its business–succeeds entirely by virtue of its invisibility and the way it reduces rather than amplifies interaction: if the very first thing you see on the Netflix home screen is exactly what you want to watch, Netflix’s AI has done its job.
In this case, AI has succeeded in Netflix’s goal of directing people toward relevant content, which is the only real reason why people pay for their service in the first place. Despite its UI being an indispensable part of the product, nobody in all likelihood buys a Netflix subscription because they enjoy hanging out in the UI, flipping through box art without ever watching anything. Netflix’s AI is valuable because it’s invisible; if it somehow interrupted you–remember Microsoft Word’s Clippy?–you would get annoyed and perhaps consider joining another streaming service.

When invisibility is a liability

However, it’s a mistake to think that all applications of AI should aim for Netflix’s model of invisibility. The way AI shapes people’s experience of Netflix is surprisingly atypical: Invisibility works for Netflix primarily because the vast majority of time enjoyed with the service is spent watching long-form content during which the AI-supported interface is completely irrelevant. While AI is an indispensable part of Netflix’s service, it is not fundamental to the most important part of the experience: watching TV and movies.
Contrast that to YouTube, which has an AI-based algorithm that may guide people to two-dozen or more pieces of content. In a fundamental way, YouTube’s AI is its experience. And because the content can be anything–unlike a TV network such as HBO with stringent editorial controls, for which the quality and tenor of their programming is a fundamental part of their brand–YouTube’s AI is its brand.
YouTube has come under significant criticism for the way the invisible hand of its recommendation algorithm can amplify hate and ignorance–watch a single white supremacist, anti-vaccine, or chemtrail conspiracy video and you get sent down a rabbit-hole of racist or baseless anti-science propaganda, all on auto-play. YouTube’s AI-based algorithm amplifies the message of whatever you watch. The dangers for young people, for whom YouTube is an indispensable educational tool, are especially acute. Is this the experience that YouTube, which in its brand mission claims that its goal is to “give everyone a voice and show them the world,” wants people to associate with its brand? Does “everyone” really mean Nazis, too?
In this context, you can see how the invisible aspect of YouTube’s AI is more of a liability than an asset. The same could be true for many other companies. Unless they make AI’s positive potential part of their brand strategy, they risk being associated with the alienating risks of AI.

Looking ahead

So how should companies brand AI? For some companies, it could mean embodying AI-generated experiences in unique personalities (think: more C-3PO or R2-D2 than HAL). For others, it might mean actively marking AI-enhanced experiences as such so that a consumer understands their behavior and benefit–and can take manual control when she doesn’t like what she sees. For still other companies, perhaps branding is a matter of empowering people to be active participants in the machine learning used to train AI systems. Imagine a service that lets consumers retrain an AI-based system when they get undesirable results.
Branding AI is a formidable design challenge, and there is no one-size-fits-all solution; as YouTube and Nextflix demonstrate, companies use AI in myriad different ways. But given the newness of AI, and the extent to which it is poised to reshape society, design-led companies should consider its branding among their top priorities.

0 comments:

Post a Comment