The terms Artificial Intelligence and Automation are often used interchangeably. They’re are associated with software or physical robots and other machines that allow us to operate more efficiently and effectively — whether it’s a mechanical construct piecing together a car or sending a follow-up email the day after your customer hasn’t completed his order.
However, the are pretty big differences between complexity level of both systems. Automation is basically making a hardware or software that is capable of doing things automatically — without human intervention.
Artificial Intelligence, however, is a science and engineering of making intelligent machines (according to John McCarthy, person who coined this term). AI is all about trying to make machines or software mimic, and eventually supersede human behaviour and intelligence.
Automation can or can not be based on Artificial Intelligence. Industrial automation can be provided by placing some sensors and making something do corresponding to sensor readings. The practice of automation evolved into what we know today between the first and third industrial revolution — production with automatic testing and control systems, mechanical labour, operating equipment and of course computers. All of the expressions of automation that have manifested around us are bound by explicit programming and rules.
If you want to make the same thing to be an AI, then you need to power it up with data. Huge quantities of data. Like use of neural networks, graphs, Machine Learning (Deep Learning) in your software. Your coding level will decide to how much extend you can make your system stimulate like human, but you are most likely going to end up at teaching the system only what you already know. In case of simple automatic you can easily predict the output, according to sensor readings. While in case of AI there is always bit of uncertainty, just like human brain.
Boston Dynamics robots constantly process large quantities of data from multiple sensors, trying to mimic human’s behaviour.
Probably one of the the biggest contribution software automation made to the improvement of the economy was the financial industry. (When it comes to industrial automation, automotive takes the lead).
In PayPal’s earliest days, fraud threatened to topple the peer-to-peer payment giant. Credit card chargebacks were soaring, criminals were using the company to launder money, and phishing attacks led to outright account theft. By the early 2000s, the fraud rate had soared above 120 basis points — costing the company millions and threatening to break already brittle relationships with credit card associations. In fact, according to The PayPal Wars, the company was once incurring $2,300 in fraud losses every hour.
To fight fraud, Max Levchin and David Gausebeck worked to develop a automated mechanism that would complicate account creation for fraud rings without discouraging potential customers or reducing conversions. What they came up with was an early version of CAPTCHA technology to block spammers from creating fake accounts.
As technology progressed, the company was implementing machine learning systems that, presented with a database of information about credit card transactions, such as date, time, merchant, merchant location, price and whether the transaction was legitimate or fraudulent, learned patterns that could predict more fraud. The more transaction data it processed, the better its predictions were expected to become, to the point where it can now predict situations just before they actually happen.
The TL;DR explanation goes something like this:
Automation is software that follows pre-programmed rules.
Artificial intelligence is designed to simulate human thinking.
But it’s a lot deeper than that — and it’s worth looking at, given that technology manufacturer Arm state in their AI survey that:
A quarter of Siri users didn’t know that AI powered the personal assistant, and awareness of AI-style tech driving other popular apps (e.g. Facebook and Netflix) is even lower. This tells us that the public’s grasp of what AI can deliver is still building.
Automated systems are everywhere. They are the reason local banks record your payments in the matter of seconds, the reason why businesses don’t have to copy-and-paste millions of personalised marketing emails to their customers. It’s what allows you to get your purchase shipped and delivered same day within 4h time slot.
Automation has a single purpose: To let machines perform repetitive, monotonous tasks or as some people say “to take the robot out a human”. This frees up time for people to focus on more important, creative tasks that require the personal touch and judgement. The end result is a more efficient, cost-effective business and a more productive workforce. An obedient digital robot that never calls in sick or takes a holiday and always gets the job done, hence, it’s no wonder that businesses so readily embrace automation.
At the end of the day, the big difference here is that automated machines are all driven by the manual configuration — which is just a fancy way of saying, you have to set up the way you want your automated system to work using workflows, programming edge case scenarios and the like. Essentially, it’s a machine that’s smart enough to follow orders.
Tech leaders have been debating these two vastly different future scenarios with AI omnipresent. Tesla’s Elon Musk — a futurologist and visionary — talks about artificial intelligence in a way that definitely falls on the dystopian spectrum saying that “Robots and AI will be able to do everything better than us, creating the biggest risk that we face as a civilisation”. Stephen Hawking was of the similar opinion when saying “The development of full artificial intelligence could spell the end of the human race.” Then we have strong AI advocates who suggest that AI will help humans but doesn’t control or intrude on their lives.That’s the whole point of AI: To create technologies that ably mimic what a human can say, think and do, which naturally won’t be affected by natural fragility (humans age and die).
And, just like most humans, that means AI is intensely bad at simply following orders. That’s not what it’s designed to do; it’s designed to constantly seek patterns (like humans), learn from experience (like humans) and self-select the appropriate responses in situations based on that (like humans).
So, what we’re actually dealing with here isn’t a simple replica of me or you. It’s about creating a system that’s more powerful than we can imagine.
Robots With Benefits
What drives both automated systems and AI is the same thing that drives businesses: data. Companies which are automated perform better and achieve substantial revenue growth. This, of course, may be a result of many factors, including the usual cross-section of benefits associated with automation. Increased productivity; better business efficiency, and most important — employees able to focus on creative or/and strategic operations.
The ideal scenario which humanity is working towards right now is this: Automated machines collate data — AI systems “understand” it. We’re looking at two very different systems that perfectly complement each other.
We’re passed the stage of recording mouse clicks and replaying it in automated way with pure RPA. But we are not yet AI ready. We’re somewhere between, at the maturing stage of cognitive automation.
If we keep investing in smart automation by fuelling it huge amounts of data, we can become much more powerful— as individuals, as businesses and as a species.