AI is not the solution to everything (or really much of anything)

Having blown away some of the hype around new technologies recently, notably AI and 5G, it is with sadness that we read that AI implementations will grow at an astonishing 162% CAGR until 2022.

AI – which we have said a number of times, is quite good at some things and quite bad at others – is on a roll. It would be a safe enough bet to say that it will follow the path – hype-wise – of IMS.

IMS, ten years ago, was the buzzword of the day. Conferences sprang up, research was abundant and peer pressure was prevalent. A survey among operators at the time concluded that IMS was ‘in our 18 month strategic plan, but I don’t know why’. Until this year, there was no clear beneficial use for IMS.

AI is becoming the answer, whatever the question.

And yet the evidence is clear enough that AI is not good enough, yet, to be the answer, except in fairly simple cases. Google and Facebook, both under enormous (and understandable) pressure to eliminate vile content from their sites, first muttered something about AI under their breath. And then Facebook went out and hired thousands of extra people to filter the content.

Let us hope that companies think this through. AI is great at looking things up and selecting an answer from a list of options. It is great at automating mundane tasks. Customer service will benefit from the technology by being able to narrow down the increasingly complex questions coming into call centres. As operators start rolling out sophisticated bundles, third party products and their own OTT-type services, they will need to answer questions that go far beyond “my broadband is playing up”, or “I forgot my password”, or “Why, oh, why do you not allow me to put hyphens in my password”.

It is also possible that security will benefit from the technology. Looking for patterns and reacting very fast indeed is ideally suited to AI.

But the hype would suggest that it can do more than that – that it can be more intelligent than us.

Let us hope that common sense precedes the wholesale and out of control implementation of this technology, and let us hope it is not seen as the silver bullet of this decade. Maybe the next decade.