Telecoms & Broadband Business Newsletter - August 2015

Editorial:- AI – don’t expect a revolution

The media loves to hype up developments such as AI (artificial intelligence), and spout doom and gloom when they talk about robots replacing humans and their brains starting to dominate ours.

I would say – not so fast. If we look at the level of intelligence that is currently on display and see how difficult it is for your PC to detect your printer and get it to work properly it is safe to say that quite a few more magnitudes of intelligence will be needed before either the good or bad things relating to AI happen.

Obviously with the current ‘big data’ developments the focus is back on ‘Big Brother’. But here as well real AI developments are few and far between. At this stage less than 20% of the companies that are involved in these developments have a good grip on their data, and even fewer are producing worthwhile results from it. The reason for this is that in the end the data in itself is not intelligent – it is up to the humans involved to make intelligent use of that data.

And for that to happen we need to take some steps backward, first formulating what questions we need to ask, so that big data can provide relevant answers. The problem here is that we often don’t know what it is that we don’t know, and asking questions to find that out is rather difficult.

It is here that our academic colleagues can play a key role. What seems to be the case is that it is sometimes better not to ask questions but to look at many different data sets and use our human intelligence to detect useful correlations and combinations. But at the same time, on a more micro level and in particular in smaller commercial environments, understanding what we need to know needs to be discussed and decided internally, and it is becoming increasingly clear that this is not an easy task at all.

In other words, AI’s intelligence totally depends on what we put into it.

On a higher strategic, and even political level, however, what is important is that we plan on that higher level for these developments before they happen – certainly not in any detail, as that would be totally futile and impossible, but in a broader sense it is possible to carry out theoretical scenario planning.

This is critical, as it is an area in which we envisage problems. The technology is moving very fast and is mainly driven by commercial organisations that want to gain a competitive advantage in the market, and, as we are seeing with some of the early developments with Google, Apple, Amazon and Facebook, this could easily lead to monopolistic situations. Around the world governments and regulators are now facing the difficult situation of managing these developments and trying to unscramble some of the technical, financial and taxation situations that have been created. Particularly in Europe we now see governments grappling, for example, with Google’s dominance in the search market and its use of private data in an ‘intelligent’ way. I won’t call this AI but it certainly is one step on that road.

A key element in looking towards future developments is that we need to try to pre-empt where the private interests conflict with the public good.

Obviously these innovative developments should be left alone as much as possible, but at the same time society has to develop, in parallel, some sort of framework that allows it to make the right decisions regarding the direction that these new developments are going to take.

While all of this depends on humans first putting the intelligence into these systems, it is crucial to be aware of the potential impact; and this in turn makes it important to make sure that we are putting the kind of intelligence into the systems that benefits human society. This is a two-edged sword. Like anything else AI can be used for bad as well as for good, and so we must stay ahead of the game to ensure that these innovations and developments are not misused.