Artificial intelligence (AI), as it relates to business, is a term used to describe a field of technologies and practical applications that give machines the ability to think and learn.
In practice, this means that as well as allowing them to do the “heavy lifting” of business and industry – as they have done since the industrial revolution – organizations can increasingly rely on machines to do the “brain work” too – thinking, planning and decision-making.
Over the last five to ten years, this technology has developed in leaps and bounds. This is mainly down to two reasons. The first is the spread of the internet into every area of business and everyday life, meaning that information is increasingly digitized and available to machines.
The second is the advances in techniques that allow computers to analyse, understand and work with this information – the development of technologies such as machine learning, deep learning and artificial neural networks. Through these, machines become capable of reading, understanding and learning from the vast mountains of data being generated and stored every day.
The AI revolution is now well underway. In finance, marketing, medicine and manufacturing, machines are learning to monitor and adapt to real-world inputs in order to operate more efficiently, without human intervention. In our everyday lives, AI kicks in whenever we search the internet, shop online or settle down on the sofa to watch Netflix or listen to Spotify. At this point, it’s safe to say that AI is no longer the preserve of science fiction, but has already changed our world in a huge number of different ways.
So: what next? Well, the revolution is showing no signs of slowing down. Research indicates that businesses, encouraged by the initial results they have seen, are now planning on stepping up investment and deployment of AI.
One of the most noticeable advances will be the ongoing “democratization” of AI. What this means, put simply, is that AI-enabled business tools will increasingly become available to all of us, no matter what jobs we do.
In the early days, benefiting from this technology would require advanced computer skills, and possibly a PhD in data science, in order to build the tools we needed as well as to gather and prepare the data with which to train machine learning algorithms. These days, due to the growing interest of business and industry in AI, tools are being made available that can be operated by anyone with a basic understanding of business IT.
As a very simple example, both Google and Facebook offer online advertising tools that rely on AI to put marketing messages in front of an audience which is likely to be receptive to them. If you work in marketing, it’s likely that you are already using these.
Likewise, if you work in healthcare, or manufacturing, or the legal profession, tools built on the same principle – generating data-driven insights and algorithmically suggesting the most efficient course of action – will increasingly become a part of your everyday toolkit.
This will be hugely beneficial for businesses as well as people. For example, rather than hiring a marketing department and hiring a team of computer experts, existing workforces can be trained and educated in the advantages of switching to AI-augmented working practices.
You’ll no longer need to be an expert in computer science to use AI to do your job efficiently – this is the “democratization” of AI and it’s a trend which will impact more and more businesses going forward.
Another trend is the movement towards organization-wide deployment of AI. The first few years of the AI revolution saw self-teaching technology rolled out into pilot schemes and small, agile deployments where lessons could be quickly learned – either through generating quick, easy wins, or by “failing fast”.
Larger organizations budgeting millions of dollars into driving these innovations are now moving past that phase, and the next step is to roll out what has been learned across their entire business operations. Armed with a greater understanding of where AI can bring benefits – and just as importantly, where it is unlikely to – we will see a move by more companies towards becoming truly “AI-first”. Effectively this means that every company strives to become a tech company – a Google or a Facebook, driven and directed by the possibilities opened up to them by innovation.
The most technologically adept businesses will follow the path set by digital innovators like Uber and Netflix – where their entire business model is built on the data they generate, and the value that it can add to their services, for their customers.
More so even that their algorithms (which can always be replicated and improved on, if the competition is clever enough) their data will be what separates the winners from the runners-up – how well they understand their customers and the ways in which they want to consume their products and services.
Uber serves as a good example here. Anyone with some training and specialist knowledge can build an algorithm which will route the closest taxi to passengers in need of a ride, but by gathering data on millions of journeys across hundreds of cities, Uber is able to understand where customers actually do want to go – as well as when they want to travel, and even why they are making particular journeys.
The third future trend in AI I want to cover here is a move towards more openness and transparency in AI systems.
This is where businesses (and in some cases governments) will come up with solutions to what is often referred to as the “black box” problem of AI. By necessity, the workings of machine learning or deep neural net-based systems are simply too complex to be easily understood by humans. To compound this, many organizations clearly have a commercial interest in keeping exact details of how their systems work away from their competition.
Unfortunately, this leads to situations where the ability of businesses to adopt AI is hindered – primarily by a lack of trust. If professionals don’t really understand how the technology works, they will be less likely to use it in their work.
An even bigger hindering factor is that if the public – consumers – don’t trust it, they are less likely to be willing to provide AI systems with the data – their personal information – which AI needs to operate most efficiently.
We have already seen this happening with public backlashes against AI-driven operations such as Facebook when there has been a perception that they’ve been handling personal data in a less-than-transparent manner.
Which is why the next step for many organizations looking to implement AI is likely to be to make themselves more transparent and accountable. Government legislation such as the European Union’s General Data Protection Legislation may force their hand in this. Increasingly, however, businesses are likely to take it on themselves to present a more open approach to their AI strategy, and those that do are likely to reap the rewards.