How to introduce AI into your organisation
In this the final of three interview with global thought-leader Ray Wang, we discuss the opportunities surrounding AI, and put some of the myths attached to this topic to rest…
Tech.Rev: AI seems to be coming out of the ‘dark space’ it has traditionally inhabited. Our fears about our personal futures are being diluted in favour of the exciting opportunities that more organizations now see for an “AI-enabled future”. What’s the reality?
Ray Wang: Artificial intelligence is something that was big in the 1990s, became a dirty word in the 2000s, and is now enjoying its proper place in strategic thinking.The big change has been cloud computing, the ability to get scalable computing power, the ability to be able to take that and process it.
The algorithms have been around for a while, and that hasn’t changed. The massive data sets? That’s also a big piece that’s actually come into play. When you pair that to the advancements made where AI is now the new UX, we can actually see things, we can translate, we can hear, we can communicate.
Tech.Rev: Who will use AI? Who will buy it?
Ray Wang: When we talk about affordability of AI people have visions of having to buy data centres, or to build digital infrastructure.That might be true in some cases, but because of the cloud we have access to compute power, and those workloads that have been put into the cloud are unique. We’ve got separate types of workloads for separate types of AI opportunities, and I think that’s what’s emerging: The opportunity for organizations to access compute power. And ultimately we’re going to pay for AI by the kilowatt per hour.
The beauty of where we are with artificial intelligence today is that the machine learning algorithms on the back end can be accessed by anybody. You could be a small startup putting a workload in the cloud, you could be a large corporation testing out an algorithm. All that is possible, because the cloud makes it easy to access the private workloads. If people want to build their own AI, AI data centres are also possible.
Tech.Rev: How fast should organizations plan to consider or adopt AI?
Ray Wang: Organizations do not have a choice about this AI bandwagon. We are seeing massive disruption and this is happening because the organizations that have built the right data collection, the right machine learning, the neural nets that are in place, are going to have a significant advantage.
This is not a digital divide. We are in a winner-takes-all-market and if you don’t have annAI programme in place, if you don’t start the beginnings of collecting that data and really be able to build that asset of finding insights, you’re not going to make it.
And it’s the organizations that have that innovative spirit, that are looking to transform their business models, that are succeeding. In the public sector, you have organizations that have built security systems and apparatus to actually find potential terrorist threats. In the private sector, organizations are improving supply chains in manufacturing.
Tech.Rev: Take us through the steps organizations typically need to take to be successful at adopting AI.
Ray Wang: There are a number elements that are important for AI.
The first is the ability to have compute power.
The second is large corpuses of data, so that organizations can mine it, look at it, look for the insights.
Third is the ability to compress time, and if you can have an ability to compress that time you now have the ability to actually make things move faster.
The fourth part is really about having the ability to think about domain knowledge. How do you apply that domain knowledge, how do you put that into a process so that you’re successful.
And then fifth, great maths talent. You’ve got to have the algorithms, you have to have the data scientists, you have to build, craft those algorithms so that they can learn from themselves and then what we need is the ability for AI to be the new UX. How do we actually engage, how do we connect. How do we actually converse? How do you take that next best action?
Not one organization has all these elements, so there is going to be a lot of partnerships around co-innovation and co-creation.
Tech.Rev: What might the near AI future look like?
Ray Wang: Let me give you an example first of what I don’t think will happen! In the early 1980s, I spent some summers in some institutions and these folks were working on big projects on how to recognise the letter A. And it took them nine months to figure out how to recognize the letter A for a computer. When I asked, “What are you going to do next?” the answer was, “I’m going to go after lowercase a.” And even then, I thought that was horrible!
That’s not happening in the future – because organizations willing to open up the new AI knowledge will leapfrog those who do not. Having said that, if you don’t keep your mission-
critical data or if you don’t control how people have access to that mission-critical data, you will be giving away your competitive advantage.
Tech.Rev: I guess we have to ask the question: what about the future of work?
Ray Wang: Some still have a dystopian view of where jobs are going to go. What we can say is when we think about factors in terms of jobs, where AI goes, one level is the volume of work. Things that are more than a human can possibly imagine, that’s going to go away. When we think about location, if things are location specific, like a plumber’s job, they aren’t going to disappear.
And then there’s also a notion of the number of connections you need for what you do. Connecting to 50 to 60 other people is still manageable for a human, but when we’re starting to work with or compete with thousands and hundreds of thousands of networks of nodes, that’s definitely going to computers and automation.
When we think of artificial intelligence, we have to think about how we’re augmenting humanity. That is the best way to understand what AI is about to do, using machine learning to apply lessons learnt and to make modifications accordingly. And that is a level of automation with a level of intelligence, that’s what’s driving artificial intelligence.
Tech.Rev: Give us some examples of AI in use.
Ray Wang: AI gives organizations the opportunity to understand demand signals, why people are making certain decisions, how those decisions are being made in what type of timeframe and what the attributes are that drive someone to take action.
AI is being used to retain employees, to identify the top employees and talents and keep them from leaving or going somewhere else.
AI is also being used to help in the area of recall. Imagine, if an eight week recall costs you
$8 billion dollars, that you could get that down to eight days:That would be $800 million. And if you get that down to eight hours, we’re talking about an $80 million recall. This happened to a large cellphone manufacturer. The ability to drive that down, prevent issues, meet regulatory compliance -these are all areas where AI has a huge opportunity.
We need to free our minds and think about what questions we always wanted to ask, that we couldn’t ask because we might not have had the data, we might not have had the maths. That’s the power of AI. It’s about opening that possibility of action questions or unlocking answers to questions we’ve never asked before. That’s the requirement for success in AI, asking better questions.