by Timur Hassan
So, how are we eventually going to deliver customised messages to individual consumers? Is it even possible to offer a one-on-one interaction with a knowledgeable sales rep, 24 hours a day, seven days a week?
There are rapid and exciting technological developments in artificial intelligence (AI) that bring us closer to this level of engagement.
Firstly, investment in real-time capable AI technology is huge. In August 2017, Microsoft announced Project Brainwave: a cloud-based AI platform which is said to process requests as fast as it receives them.
But if you want to keep AI in your own hands, graphics card manufacturing giant nVidia has created hardware called the nVidia DGX-1, which can process AI algorithms many times faster than even the most powerful servers. While these technologies provide the power required to create more and more advanced AI, great leaps have also been made in the algorithms that make AI possible.
In March 2017, Google’s artificially-intelligent bot AlphaGo beat the world's best player of the complex board game Go in four games out of five. This happened many years sooner than experts predicted it would, as Go is extremely complicated and allows for an unimaginable number of possible moves.
This achievement was surpassed in August. Elon Musk's OpenAI learned, after only two weeks of practice time, to play the computer game DOTA 2 to such a high standard that it beat a professional player in both of the games it played. This game is one of the most popular and competitive video games of all time, breaking the world record in 2015 for the largest prize pool in eSports’ history: a whopping $18 million. The professionals that have played DOTA 2 against OpenAI have been stunned by its capability.
Also in August of this year, Microsoft broke the record for human speech recognition with an error rate of 5.1% – on par with the error rate of humans and professional transcribers – in a standardised test.
As well as being able to understand what you’re saying, AI – in the incarnation of IBM’s Watson Personality Insights – has also been used to analyse personalities. IBM Watson does this through the written word, and measures psychological profiles in terms of the ‘Big Five’ personality traits: openness, conscientiousness, extroversion, agreeableness and neuroticism (OCEAN). This widely-accepted model is used in psychology and marketing to explore how personality traits correspond with and predict various behaviours.
So, the hardware used to process AI algorithms is developing at a mighty rate, and AI is matching or beating humans in complex tasks, including speech recognition. Simultaneously, we’re seeing robots that learn about your personality as you communicate with them.
How long will it be until we have a 24/7 AI sales rep that speaks to you through your favourite messaging platform or via a phone call; understands your psychology; and learns, through experience with you and others, the kinds of messages that are most likely to inspire you to act or buy, based on your personality? It may not be long at all, and that would be a revolution.