It’s hard to say. Smart machines and robots are becoming increasingly common in the workforce, but humans won’t be side-lined any time soon. As they say, know your enemy.
First of all, the big question: what do we mean by AI?
Artificial Intelligence (AI) is a branch of computer science that deals with creating systems that display intelligence. Typically, most of us think of Skynet from the Terminator movies, but AI is just as commonly used to describe narrower, but still very impressive, “Applied AI” – for example, a computer program that plays chess or detects credit card fraud. So far, the most common way to develop AI is through what’s known as Machine Learning.
Go on.
Machine Learning (ML) is a field of AI that aims to teach computers to learn to extract meaningful insights from data sources and perform a task without being explicitly programmed to do it. Key data sources include Natural Language Processing (NLP), Computer Vision, Audio Processing (like speech recognition), Time Series (like stock forecasting), and Graph Analysis. If you want to know more, have a look at Distil, a journal dedicated to a clear explanation of ML.
Easy: it’s a computer that learns.
The most important ML technique is called ‘reinforcement learning’. Inspired by behavioural psychology, it focuses on how software agents should take action in an environment to maximise a reward. It’s like software training at the gym. This affinity with human behaviour continues with neural network. These are designed to simulate biological networks, as in pattern recognition, language processing, and problem solving. The goal? To self-direct information processing. Note: when neural networks have numerous layers it’s called ‘deep learning’. Go deep in deep learning with Google zero-shot translation story.
Wow! So what can a learning machine do?
Pretty much everything, starting from automating repetitive tasks usually done by humans – especially if you combine so-called Robotic Process Automation (RPA) with innovative smart engines to create Data Robotics. This is a new level of process automation based on self-learning technologies and AI. It aims to improve productivity and efficiency by automating activities and procedures as never before! A Data Robot is the software that can analyse situations, understand complex information, and learn and operate in a way that optimises and supports human activity. Following a learn-by-error approach, it can learn how to carry out complex transactions, manipulate data and improve system integration.
Robots sorting system helps Chinese company finish at least 200,000 packages a day in the warehouse.
Help! I’m afraid a piece of software is going to steal my job!
Don’t worry, we’ll still need the human factor for numerous job for years to come, but machines are now rapidly becoming part of the workforce. Is your job robot-proof? There’s only one way to be sure: Lifelong Learning Education, as done by 50 Replyers and customers in Trieste past February to develop the first Data Robotics projects in the dedicated Reply Hackathon. So, be curious and keep reading Reply’s R20!
Reply Hackathon on Data Robotics, held in Trieste on February 2017.