Lets start with Where is AI ?

According to Paul Clarke, CTO at Ocado, “AI will be everywhere and in everything”.

Understanding where AI is – in terms of where its being used, and where it fits in relation to other domains  is a crucial first step to understanding what it is and how we can use it.

But before asking what or where Artificial Intelligence is, it’s worth asking what intelligence is. A commonly used definition of intelligence is “goal-directed adaptive behaviour” Sternberg & Salter[i].

Artificial versions of intelligence are certainly goal-directed, but as yet AI can’t be considered to be fully adaptive.

A more helpful definition states that AI is a branch of computer science dealing with the simulation of intelligent behaviour in computers.

The fundamental processes behind AI can be broken down to just three key steps –

1)    Preparing data (which is between 60 - 80% of AI work)

2)    Applying statistical algorithms to that data

3)    Obtaining and evaluating predictions

To appreciate AI more fully, we need to explore the Russell Ackoff data-to-wisdom continuum where:

·      Data is symbols

·      Information is data that is processed to be useful and provide answers to basic questions

·      Knowledge is the application of data and information to ‘how’ questions

·      Understanding is appreciation of "why"

·      Wisdom is evaluated understanding

AI can be used to extract knowledge from information and data, whilst human domain specialists can turn that knowledge into wisdom.

As with most technologies, it’s a reasonable expectation that AI will become commodified, and that general-purpose AI tools for business will become widely used. In 1962 the concept of a spreadsheet was embodied in Fortran[ii] - a language accessible only to specialists. By 1985, the spreadsheet had become fully visual in the form of Excel for the Macintosh. Now Excel is ubiquitous. It’s not unreasonable to expect a similar journey with AI. However, for now, and into the foreseeable future, AI is complex, messy and requires a lot of intellectual horsepower to design, build and successfully implement.

AI can be thought of as being nested within a number of domains, and AI itself contains a number of sub-domains.

AI is part of the broader domain of Big Data, which itself is a subset of the wider domain of Data. Within the AI domain, is machine learning, and within machine learning is the domain of Deep Learning.

[i] http://www.indiana.edu/~cnilab/intelligence.pdf

[ii] https://en.wikipedia.org/wiki/Spreadsheet#Batch_spreadsheet_report_generator

Figure 9. Where is AI?

AI is moving forward rapidly, so it’s worth thinking about its evolution. Right now, AI is limited to narrow tasks, but the goal is to make AI that is multipurpose. “No computer program today can match human general intelligence,” says Murray Shanahan, Professor of Cognitive Robotics for the Department of Computing at Imperial College in London. “Humans learn to achieve many different types of goals in a huge variety of environments. We don’t yet know how to endow computers with the kind of common sense understanding of the everyday world that underpins human general intelligence, although I’m sure we will succeed in doing this one day.”[i] 

Broadly speaking, AI applied to a specific task is called ‘Narrow AI’. ‘General AI’, on the other hand, is where the same AI system can be applied to any problem, and we are some way off from developing this. As AI evolves, a useful way to look at different types of AI is: 

Assisted intelligence - Automating repetitive, standardised or time-consuming tasks. This is happening now, and fairly commonplace

Augmented intelligence - Humans and machines collaborating to make decisions. This is happening now.

Autonomous intelligence - Adaptive continuous intelligent systems take over significant amounts of human decision making. This will probably happen in the future.

The practical reality is that true artificial intelligence is a long way off, but the term AI is too widely used to make correcting the terminology worthwhile.

In practice, what we encounter is machine learning.

We can think of this of today’s AI as part of the attempt to reach true artificial intelligence, and more specifically of computer science, which deals with the study of systems and algorithms that can learn from data, synthesizing new knowledge from them.

The word ‘learn’ intuitively suggests that a system based on machine learning, may, on the basis of the observation of previously processed data, improve its knowledge in order to achieve better results in the future, or provide output closer to the desired output for that particular system.

The ability of a program or a system based on machine learning to improve its performance in a particular task, thanks to past experience, is strongly linked to its ability to recognize patterns in the data. This theme, called pattern recognition, is therefore of vital importance and of increasing interest in the context of artificial intelligence; it is the basis of all machine learning techniques.

[i] http://time.com/4960778/computers-smarter-than-humans/