December 20198 ARTIFICIAL INTELLIGENCE ISN'T HERE TO STEAL YOUR JOBBY TIM WARREN, ANGEL INVESTOR, FLYING KIWI ANGELSThe way we work is changing at a rapid pace portfolio work, `slash' workers, fractionalised jobs and the influencer movement. While the concepts have been around for eternity, these are now widely accepted approaches in the last decade.Who would have believed 10 years ago that 25 year olds would be paid to travel around the world and take photos of themselves? In 2009 who would have thought that you could efficiently be a private driver to hundreds of strangers, or hail a cab effortlessly in hundreds of cities.Yet these are just the beginning stages of a technology revolution that is redefining the future of work.The real revolution comes from machines, computers, phones that simulate human intelligence and start to do the lower end of our daily tasks - freeing us up for more challenging activities. With the advent of vast stores of computing power provided through cloud systems, AI platforms are arising that enable computers to do truly human like tasks. There are a range of tasks that are more suitable than others for this early automation. Those tasks that are high volume and repetitive, process driven and well-defined; these are logical places to first apply automation. One must distinguish between simple automation and artificial intelligence in this context. A process that simply takes a predefined input, performs a task and delivers an output in a way that does not involve humans is automation. Steady advances in automation have been occurring since the beginning of the industrial era. This is not what we're talking about. Automation through artificial intelligence is an entirely different field. While the objectives are comparable: more work in less time, the methods and the capability are significantly different.Artificial intelligence based approaches simulate human intelligence through statistical methods so that inputs can be varied and non-deterministic.Let's look at a real example.A pattern matching automation needs the exact phrase "I want to buy a car". If a user says "I need to buy a car" , the simple process will fail. Artificial intelligence approach takes multiple training phrases. "I want to buy a car", "I need to buy a car", "Sell me a car", "I need a car", "How do I buy a car", "Get me some wheels"... etcetera, you get the picture...From here it builds a statistical model and anything that the user asked for is compared with that model resulting in a percentage confidence that the user actually wants to buy IN MY V EW
< Page 7 | Page 9 >