AI should not be a solution looking for a problem – we should not be thinking ‘How do we use AI in our business?
The date for technological ‘singularity’, the hypothetical advent of artificial intelligence (AI) or the point at which this technology will be better at redesigning itself than humans will be at creating it, is unclear.
The term was coined by mathematician and physicist John von Neumann in 1958 and popularised by computer scientist and science-fiction writer Vernor Vinge, who believes singularity will occur before 2030. The futurist Ray Kurtzweil sets the date around 2045.
With most predictions falling in the next 15 to 30 years, it is no wonder that AI is becoming a dominant theme in popular culture. Major films such as Ex Machina, Lucy and Her, as well as Channel 4’s series Humans, all focus on AI, approaching it as very much a human and emotional story. In a conversation with sentient robot Bina48 this month, Marketing also discovered the human side to robots first hand (see videos below).
If technology can enhance intelligence and emotion concurrently, then AI is an interesting prospect for marketers.
Nonetheless, it brings with it serious ramifications. Indeed, media agency PHD took the topic
to this year’s Cannes Lions, publishing a book on sentience and marketing and inviting Sir Tim Berners-Lee, inventor of the worldwide web, to talk on the subject.
However, AI is already being used by major brands – BMW employed the technology two years ago in an ad campaign for its first electric car.
Its iGenius technology was able to answer customers’ questions about the new model via text, reducing the need for BMW to invest in training dealers or customer-service staff to handle queries.
Real...but no too real
Stuart Walker, the head of BMW brand communications, says: "Not only was iGenius able to have hundreds of conversations at the same time, it could also memorise previous questions from each person to generate answers that were meaningful."
However, he warns that the technology needs to be both agreeable and accurate for people to be comfortable with it. "It was important that iGenius was set up with personable responses so people did not feel like they were talking to a machine," says Walker.
"BMW did not face any issues. However, it was important that we trained the AI to have the same exacting standards and level of knowledge as any one of our human experts, if not more. A lot of time went into putting information into the tool and ensuring it was up to date with our latest offering."
Retrofitting AI onto static advertising in this way was a smart move. BMW’s media agency, Vizeum, has since used the technology for other clients, such as Panasonic, to enhance sales and improve the effectiveness of otherwise non-interactive in-store experiences.
Most people are still wary of AI – if it feels too human, it will fall short.
Chrissy Totty, head of innovation at Vizeum, says: "Most people are still wary of AI – if it feels too human, it will fall short. Be transparent and clear about what you are doing and don’t try to pass it off as a real human being."
But for software and services company Evernote, a brand that has used AI from the outset, the technology has to improve experiences for customers – to be ‘prescriptive’ and adapt to different circumstances, rather than be a ‘one-size-fits-all’ approach.
"We have made sufficient advances to enable us to use a lot of integrated intelligence in our applications," says Zeesha Currimbhoy, vice-president of augmented intelligence at Evernote.
"But I believe that AI should not be a solution looking for a problem – we should not be thinking ‘How do we use AI in our business?’ Instead, it should start with the user experience and how that can be improved with the advances we have made."
From a communications perspective, David Caygill, the creative tech director of creative agency Iris, says there is an opportunity for brands to use AI to provide frontline communication and customer service, but "only when it gets a lot better at understanding what people want.
Repeating yourself to a machine is one of the most frustrating things a human can experience."
While the advent of general AI may still be years away, brands must get a handle on these nuances now. Policy will be forged on the basis of AI’s moral and social implications, and it is already clear that brands will need to understand what this will mean for the customer experience.