The lowdown on the high expectations for AI

The lowdown on the high expectations for AI

12th February 2018 by Steve Bates

Most people have heard of Artificial Intelligence (AI), yet start a conversation about it and it soon becomes apparent that most people only have a loose understanding of what it is and the heady heights it could reach.

So, what is AI?

The name ‘artificial intelligence’ is actually a little misleading because the technology in this area doesn’t have much to do with human type thought or true human intelligence. In reality, it’s an alternative way of programming computers that uses huge quantities of data to train computers to carry out a specific task.

It involves setting a target outcome and instructing a computer program to find out what actions are needed achieve it, then carry them out – with minimal human input once the desired outcome’s been set. It’s the ‘creative’ element that makes the potential for AI so huge, despite getting off to slow start.

Let’s go back to the start: Deep Blue and the birth of AI

 Back in 1997, a computer created by IBM called Deep Blue, had a six-game chess tournament with Garry Kasparov, the then reigning champion. It became a media sensation and was billed as the ultimate man versus machine contest. Deep Blue narrowly won the contest with two wins for the IBM machine, one for Garry and three draws.

The point of the tournament wasn’t widely reported at the time. The contest was really about demonstrating the ability of a computer to handle complicated calculations. This ability could then be harnessed to carry out the complex financial modelling needed to carry out risk analysis, identify trends, and search massive databases. In short, it was about seeing if a computer could handle the huge calculations that many fields of science need.

The type of Deep Blue AI technology used wasn’t general enough to solve the type of calculations that breakthroughs in science demand, so it didn’t lead to a quick leap forward in computer programming.

Learn a little about machine learning

In technology, neural networks are sets of connected, simple calculators. There has been exciting progress in machine learning recently, in an area known as ‘deep learning’. This is where a network is arranged into multiple layers between an input, such as the pixels in a digital image and an output, such as someone’s face within the overall picture.

The network is ‘trained’ by giving it lots of inputs (in this example, a high amount of pictures that include images of human faces). To continue the example, the outputs would be the accurate identification of the people within the images. But machine learning can also be used for far more complicated procedures, for example, ‘teaching’ a driverless truck to safely drive on a busy road.

What is AI already used for?

Fast forward a couple of decades past Deep Blue to today and AI has made huge progress. Especially in the area called ‘machine learning’.  This has made AI particularly good at finding patterns in data and making predictions. In fact, it’s already being used for:

  • Search engines that rank websites by relevance to the search term
  • Speech recognition services
  • Faster decisions for financial services, such as whether or not to offer a loan or extend a credit card
  • ‘High-frequency trading’ algorithms use pre-determined decision criteria to respond to financial market conditions many times faster than human traders
  • Product recommendations for services such as Amazon and Netflix
  • Smartphones and digital assistants such as Amazon’s Alexa, that enable us to ‘talk’ to them to ask questions and receive answers
  • Driverless cars and bus prototypes
  • Companies such as Amazon use AI to plan the most efficient delivery routes and for optimising warehouse storage

And the list of potential applications for AI – and innovations it supports – is growing rapidly.

Why all the sudden excitement around AI?

Although AI is still in its infancy, it’s now much easier to see the potential it offers.

It could transform productivity and innovation. AI may lead to more prosperous economies and take care of labour intensive work – replacing it with more fulfilling careers. It could even enable brand new business models and solutions.

What’s more, AI and its applications are rapidly evolving, so there are probably business benefits that we haven’t thought of yet.

Should we be afraid of AI or welcome it?

Despite sci-fi films about robots going rogue, AI has the potential to have a huge and positive impact. It will probably be the defining technology of our age. Just as steam-powered technology was when it arrived.

But should we be afraid for our jobs? Well, there have been similar huge changes triggered by a change brought about by the introduction of technology.

For example, the change that cars made when they became available to the masses. When cars became commonplace, jobs that relied on horses went into steep decline. But whole new industries were created around the car, which led to more jobs in total, with new work in:

  • Car manufacturing
  • Road building
  • Fast-food outlets to cater for motorists
  • Hotel and motels
  • Petrol stations
  • Vehicle servicing – from oil changes to replacing tyres

So although in the short-term AI may lead to job losses, in the long-run there may be a positive effect on employment. It could enable humans to take on more creative and rewarding roles. And free-up time to spend on life outside of work.

If you found this article interesting, you may like to see our post on making a profit from investing in early-stage companies.

Steve Bates

Steve Bates

Content Team at Seedrs

Digital Agency Kent