Sophie Lord: the executive director of strategy for Landor
Sophie Lord: the executive director of strategy for Landor
A view from Sophie Lord

The five laws for branding artificial intelligence

How brands can create AIs that won't turn into Hal overnight.

The rise of artificial intelligence (AI) is perhaps the single biggest challenge facing today’s brands.

This may sound like hyperbole, but anyone familiar with the development of AI over the past few years will know how disastrous they can be for brands if they get it wrong. Microsoft’s chatbot Tay is the obvious example. She morphed from an innocent teenager to a Hitler-sympathising robot within 24 hours. Microsoft quickly shut down the project, but not before receiving a barrage of criticism.

Even some AIs considered successes create branding issues. Apple’s Siri, and Amazon’s Alexa are some of the most prominent personal assistant AIs. However, their femininity combined with their role as a house-bound helper has been interpreted as regressive stereotyping built into their design. While clearly not intentional, this implicit reflection of the sexism underpinning Silicon Valley innovation is regrettable.

So, how can brands do better when developing AIs? In the spirit of Isaac Asimov, we at Landor have created what we see as the five laws of branding AI:

1. Your AI = your brand

Any AI your business creates will become a part of your brand as a whole.

In fact, it has to. Marketers and brand managers need to recognise that automated bots are more than purely a service initiative and can directly impact a brand’s standing in the marketplace. While still relatively untrodden ground for the marketing department, AIs will act just like any other customer touchpoint and their brand implications need to be acknowledged.

Brands need to get a handle on their AI brand strategy quickly. Brand managers that do not embrace AI are bound to face a number of challenges as the market and their offerings develop. Would you be happy if your chatbot didn’t echo your brand’s identity or values? 

2. Make conscious decisions

Marketers cannot simply reflect society in their work. This approach increases the risk of a brand reflecting the uglier side of human nature, which needs to be avoided.

This is just as true when considering AI technology, especially in light of the ongoing debate around gender identities – just look at Microsoft’s Tay. Bots have suffered from issues ranging from static gender roles to racial biases concealed in AI language use. Brands that are responsible must instead take a deliberate and positive stand in the ongoing gender debate through their AI design.

And what’s the best way to achieve this? Make conscious decisions when developing your AI. This may sound counter-intuitive ("all our decisions were made consciously!") but it really means considering the full ramifications of each feature of your design on your brand, your users and society.

Conscious decision making is also incredibly important to make an AI as effective in its purpose as possible. Consider smoke alarm bot Nest Protect, which uses a female voice that actually plays an important role in its function. Research from the University of Dundee found that children respond best to human voices when warned of danger, such as their mother’s voice. As such, a conscious decision from Nest increased its AI’s effectiveness in warning a family of fire.

3. Embrace non-binary

One of the most outstanding recent developments in society has been the embrace of non-binary values. From "Call me Caitlyn" to boyband of girls Acrush, we’re finding ourselves in increasingly fluid times. Why shouldn’t brands embrace this development?

More progressive brands are already following this agenda and embracing the possibilities it presents to them. HSBC has introduced ten new gender-neutral titles to accommodate for the growing range of gender identities, winning widespread praise for the brand as a result.

If this thinking is applied to the next AI chatbot, its brand could see significant benefits as a result.

6. Go beyond human

Watson, Siri, Alexa, Cortana, Rosie, Tay – all AIs, all anthropomorphic and all designed to have human characteristics. But why this focus on human-centric robots given the evidence that we can accept and even bond with non-human technology?

Japan is well ahead of the curve on this. Paro the therapeutic seal is having a real impact on old people’s lives, while funerals are still being held for Sony AIBO robot dogs after the technology company discontinued support for the tech-pet.

By taking a leaf from Japan’s AI playbook, brands have a chance to be truly creative with their AIs.

5. Be agile

In a way, this final rule encompasses the whole of this list. When approaching your AI brand experience, have fun with it and consistently rethink how it interacts with your customers. Move from bland and stereotyped to diverse and differentiated. Avoid explicit binary thinking and embrace the challenges posed by a society in flux.

But most of all, be ready for further change. Bots and AI could soon become brands in their own right, with Alexa being viewed wholly separately to Amazon. There’s also the prospect of brands interacting directly with AIs rather than consumers – Google Home could well be choosing laundry detergent brands on consumers’ behalves.

Brands that are flexible with their AIs are most able to react to such changes as they arise. With these laws in hand, they should be ready for anything the world can throw at them or their AI creations.

Sophie Lord is the executive director of strategy for Landor.

Topics