Artificial intelligence is having quite the moment. Everyone seems to have a view, or theory, on what it is and what application it might have.
Dig a little deeper, however, and a different picture emerges. An incredible 45% of consumers think AI 'is a robot', according to a recent survey VMware commissioned. There's a perception that AI is a 'thing' rather than an intelligence built into many services and systems we already receive and use.
This confusion is understandable when you consider that few of those evangelising about the technology are tying it to real, valuable applications. It all leads to the question: when we talk about AI, what do we mean? Do we mean a system that is truly intelligent and can do whatever it puts its artificial mind to, or are we referring to a very specific function, which the system executes and then improves on? More simply, are we talking about artificial general intelligence, or task-based intelligence?
It's this confusion that means we end up going around in circles on the moral ambiguities of driverless cars or, to an extreme, AI-powered robots rising up like some sort of unstoppable terminator. Alex Champandard, co-founder of creative.AI, summarised this by explaining that AI is being used as a label to address different fears people have around technology innovation. He said: "Some of them are afraid of it going to war or afraid of it taking their jobs, but it is basically just us projecting all of our worries into some abstract concept."
The point is that people are concerned about their futures generally. The world is changing at a rapid rate, and, understandably, we all want to know what's going to happen. AI's fault is that it has the perceived potential to make those fears of robotic war, of mass unemployment, a reality.
Businesses should take note for two reasons. Firstly, VMware research shows more than half of consumers are specifically looking to organisations to help them understand what game-changing technologies like AI can actually do. Secondly, the possible applications, looking at task-based AI rather than broader general AI, are too great for businesses to ignore. The opportunity lies in balancing the pursuit of gains with the genuine concerns that lie underneath the melodrama.
Augmenting business ability
Take a step back; there's a huge amount of potential improvement technology has, is and will continue bringing to businesses. The issue is that as clouds, applications and infrastructure expand and become more sophisticated, something is always needed to manage it effectively and efficiently. To put it another way, how can we manage an increasingly intricate technology landscape at scale? Our human workforces are smart, but they have limits. Our systems have become too complex for our minds, and maybe some of the work just too 'commoditised', which is where AI, with machine learning, comes in. In the UK, for example, the National Grid is using drones to inspect its 7 200 miles of power lines and is now applying machine learning to reduce the raw footage an actual human sifts through.
Optimising processes may not sound as exciting as AI-powered robots or autonomous galactic explorers, but it is the beginning of a genuine business revolution. For business, the process is threefold.
First, the current state. Humans get reports on how their systems are running and make decisions based on this. That might mean the performance of different cloud environments, data centre optimisation or even the quality of data in a CRM system.
Next, we move on to the desired state. Humans decide what the system should look like and deploy machines to ensure this is realised. This is where the system is programmed to learn how to do something, otherwise known as machine learning. Applications that meet certain criteria are automatically repositioned into the environment best suited to meet their needs, data centres are deployed as resource demands increase or decrease, and data is automatically cleaned as it comes into the CRM.
The final point is the future state. AI comes in. It builds on what machine learning is doing to work out how the system should be configured to deliver maximum results, potentially in ways humans haven't thought about. It is this future state where things start to get really interesting, augmenting and improving, without constant human input.
Yet, we need to get there. As the UK futurist and entrepreneur Sophie Hackford recently asked, are we too stupid to be able to program AI properly?
Staying between the guidelines to achieve success
Hackford was talking about AI at a broad scale, a generalist AI, and how we may be too conflicted to be able to fix some of the big problems, such as climate change, pandemic management or food security, which we face as a race. Yet, it's a notion to be considered at a business level as well. Do we have the capabilities to deliver truly radical innovation, driven by AI? In essence, are we capable of reaching that future state?
Perhaps. Here's another consideration: is that the question we need to answer now? We talked before about the AI confusion, and that being linked to people espousing the theoretical possibilities, without linking it to the reality. Yet, actually, there's a huge amount already taking place. In our three phases, we have already reached phase two.
Gmail, for instance, uses machine learning to limit spam. Uber too; it incorporates the technology into ride ETAs and food delivery times. Chatbots in any form of online support are powered by AI. Those are just the instances we as consumers are likely to interact with AI. In data centres, Google was able to use AI to cut its energy bills by 40%. Bank of America is using AI in its intelligent virtual assistant named Erica, designed to perform day-to-day transactions for customers, as well as anticipating their individual financial needs by providing smart recommendations.
In all these instances, AI has been developed to deliver on a specific set of criteria, which is how the implementations have been successful. Like teaching a child between wrong and right, AI needs parameters, which need to be programmed by humans.
The other point to consider is that AI is ultimately a business tool, not a strategy. The focus needs to be on what the underlying business issue is, and then, if appropriate, work out how to integrate AI, just like any other piece of enterprise technology, whether it's cloud, blockchain, virtualisation or mobile working.
Speedy Hire, a UK tool hire company, recently announced improved financial results, linked to a renewed focus on small and medium-sized enterprises, promising next-day delivery on selected products, and four-hour deliveries within a certain geography. It could only do this by knowing what customers want and having the right stock in the right depot. This was achieved through the deployment of data, artificial intelligence and machine learning helping it target its resources where they are needed, along with marketing offers to new customers.
The strategy was to target SMEs with services (such as faster delivery) that would appeal to customers. AI was simply one of the tools to help it develop the solutions to be able to offer those services. That's not to say it wasn't important; without it, being able to understand what customers wanted, when they wanted it, would have been much harder. Yet it wasn't the strategy itself.
So, teach AI like a child; don't fear it like a terminator
The conversation around AI needs to change, away from concepts beyond our comprehension, and start to become rooted in actions that can have a tangible impact. It has real potential to augment business, if handled properly.
Ultimately, while it is important to be cognisant of the fears about AI and where they come from, looking at the extremes helps no one, whether it's a positive concept of an AI servant for everyone, or a negative one where said servants enslave us. Just like humans, AI does have limitations, and that's where we come in. AI is a child, it needs guidelines. Our role is to be able to understand what businesses want to achieve and see where AI can help.
Without these guidelines, AI is just another technology likely to get lost in its own hype. With them, however, it can be a powerful new element of business success.
Share