Artificial Intelligence: Why you can’t afford to ignore it any longer

By Jonathan Bowdler, Head of Regulatory Compliance, ICA

Once, public understanding of artificial intelligence (AI) derived only from science-fiction novels and cinema – various Terminators, I, Robot, Blade Runner and the latest iteration of The Matrix all testify to AI’s enduring appeal to the imagination.

Today, however, the picture is completely different. Knowledge of AI in the workplace has been growing steadily in recent years, especially the potential impact of AI systems. And whilst computers have been with us for decades, it is only now that AI has become a feasible tool capable of regular use within industries.

This means that professionals must understand the basics of how AI works, and its potential to revolutionise much of what we do.

First, we need to be clear on what AI actually is.?

Computers today possess such advanced processing powers that the speed itself can begin to give the impression of AI. Yet all that is really happening is that the computer is processing its programming so efficiently, and making decisions so quickly, that it gives the impression of intelligence. An excellent example is Deep Blue, the first chess computer to beat a human Grand Master. It simply processed the possible outcomes of every potential move throughout a game (of which there are billions) always choosing the move most likely to attain victory.

True AI, though, is best defined as the simulation of human intelligence processes by computer systems. These processes include:

? knowledge acquisition, including language recognition

? learning, e.g. identifying the ‘best’ way to follow instructions, using the results of previous actions

? reasoning, i.e. criteria-based decision making, in addition to learning

? problem solving, i.e. deciding how best to apply the result of its reasoning, and

? perception, e.g. the ability to consider the reaction of others, both human and artificial.?

True AI is actually called artificial general intelligence (AGI) and would be able to replicate the way a human brain works, no matter the scenario it is presented with. Whilst we are certainly not at that stage yet, we are starting to develop computers that can replicate a simple brain within a narrow set of parameters.

The hierarchical development of AI

Level 1: Reactive AI – very fast processing gives the appearance of intelligence. Limited scope of operation dependent upon programming and purpose.

Level 2: Machine learning – the machine ‘learns’ from its previous actions, meaning it can improve outcomes in similar scenarios. Limited in scope dependent upon programming and purpose.

Level 3: AGI – mimics the human brain. Able to consider outcomes and impacts as part of decision making. Can apply its capacities to any scenario. Likely to communicate through language.

Level 4: Awareness – AGI with a sense of ‘self’ which could be described as consciousness, or even life.

Levels 1 and 2 are very much with us now, whilst level 3 is being developed in laboratories around the globe. Level 4 will probably have to wait for quantum computing to be fully developed (to be explored in a future article).

Given its seismic potential, now is the time to begin your AI learning journey; you will likely require some knowledge whatever your role, and you will need a great deal more if/when you become involved in designing or overseeing AI systems. Fortunately, it is a fascinating subject. Just look how much science fiction is written!

If you are interested in finding out more about ICA’s learning solutions, please visit www.int-comp.org/corporate/ica-enterprise.