AI is about intelligent as your front door right now – MWC feedback

And we’re certainly not talking about one with a smart lock. Artificial intelligence could prove to be a game changer for those who nail it, don’t believe the hype just yet; AI is not intelligent.

After numerous conversations focusing exclusively on artificial intelligence we had several conversations which said nothing new, and several which confirmed there isn’t much to the buzz right now. Yes, AI is going to be one of those technologies which is going to change the world, but we aren’t anywhere near that stage just yet. Don’t listen to the hype or the buzz from all the vendors which are promising you riches in 2018, AI is still in the embryonic stages of development and not much use.

“We are seeing progress, and we are seeing acceleration because of the cloud, but AI is not intelligent,” said Brian Kracik, Director Communications Industry at Oracle.

Looking at the dreaded idea of Skynet, Kracik highlighted that we are so far away from this being a reality it isn’t even comprehensible. Right now, state of the art AI technology would be considered as leveraging data to automate decisions, but when you look at the greater potential of AI this is simplistic. It is taking data which says A and then performing action B. This is not exactly what you would call deep seeded intelligence.

These are of course the first baby steps, and for the operators this will be into the customer call centre. We’re not necessarily talking about the end of human customer service agents, but Kracik pointed towards the idea of context. An AI assistant in the customer service knows every customer and doesn’t forget details. CRM systems have done this in the past, but anyone who has worked on one knows that they are certainly not perfect.

Intelligent customer systems which can analyse sentiment during voice and video calls, track a customer on your website and make intelligent assumptions on future behaviour. This is the first step, it is very simplistic, but as compute power increases alongside the amount of data, the capabilities of AI will accelerate.

Once this simple idea of context has been achieved, the next could be revenue generation. This is tricky though. Identifying patterns in purchasing behaviour is difficult, as it making the right recommendation, but doing it in a way which is not felt as intrusive or invasive by the consumer is a completely different story. User acceptance is one area which might hold back the technology for some time, but this should not overshadow the immaturity of the technology. Sometimes we just need to realise how far in the journey there is to travel.

Exploring the idea of fear isn’t a bad idea, and one which is currently on the mind of Ulrika Jägare, Director Analytics and Machine Intelligence at Ericsson. We not saying that AI is going to rise up and take over the world like it has in many sci-fi movies, but ethics need to be designed into the machine. Now this is a ridiculously complex area of AI.

First of all, Jägare highlighted AI is governed by policies. These policies need to hold the machine accountable to the ethical principles of humans but also to our laws. But, at the same time these policies need to be flexible enough to allow for the concept of self-learning. In a perfect world, AI would be allowed to self-govern, rewrite these policies to allow for greater learning and benefits, but then again you also have to demonstrate an element of control and accountability over the machine. Walking the tight-rope between self-optimising policies and maintaining human principles and ethics is a complicated job which we haven’t even begun to scratch the surface of.

“We haven’t got anywhere near putting the whole puzzle together yet,” said Jägare.

Of course before any of this can be achieved, the workforce needs to be completely redefined.

“Workforce transformation is a common topic for us to have discussion about,” said Emma McGuigan, Group Technology Officer for Accenture’s Communications. “It’s interesting how the organization who aren’t quite there think of AI as a technology which is going to cost jobs.”

Breaking down siloes in an organization is a tricky task, as is overcoming the fear that your job could be lost to a machine. This is the problem with initial AI discussions, it has always focused on the simple task of efficiency not adding value to an organization. The most progressive organizations and the ones which are making money in the digital economy are the ones which are searching for new revenues through technology, not just using technology as a means to save money.

Behaviour is of course the most difficult thing to change. This is not just in the business case, but in the wider world as well. For the full benefits of artificial intelligence to be realised, the approach to how technology is managed, integrated and utilised needs to be changed.

This feedback from three senior executives in the industry is not meant to be a dampener on the technology, but more of a reality check and a warning. Artificial intelligence will eventually underpin every technological advancement of the next few decades. Whether we are talking about customer engagement, autonomous vehicles, augmented reality, zero-touch networks or the greater digital business model, AI will be the foundation on which these advancements are built.

But we are miles away from this being even the most basic of realities. So be wary about anyone trying to sell an end-to-end AI solution, snake-oil salesmen can be geeks as well.