IBM says AI is going to be critical to telecoms, but the key is doing it ethically

In the first of a two part interview with technology giant IBM, Stephen Rose, IBM’s global General Manager for TME and Distribution (pictured) discusses the role of AI in the telecoms sector and beyond.

IBM is one of the biggest players in AI development, could you describe the current state of AI in terms of how it’s being implemented by telecoms firms into their operations, and what the actual benefits are?

AI is central, I think the key thing is doing it ethically, that’s number one. And anybody that understands the notion of trolleyollogy will start to understand some of the difficulties around ethical AI implementation. If you’ve got autonomous vehicles, to what extent you’re going to have localised decision making and whatnot, then there’s always going to be those issues in an emergency situation. In the immediate term, it’s pretty clear AI can be very, very well implemented across customer service environments.

AI can project towards me some of the issues that I might have proactively, so that I can then be intercepted as a client and be given workarounds. There are a whole bunch of AI decisions contained within the customer services environment. On the network side and on the IT side, I think there’s a great deal of opportunity to do closed loop automation, and if I’ve got a certain type of service, AI and automation can actually figure out to what extent the service may become interrupted. The way these systems work is that they predict based on an analytical probability, whether or not the service is likely to require intervention or not, and if it requires intervention. And that’s where AI definitely has a role.

Machine learning has a massive role in the future. If you want to offer somebody a slice, and that slice has eight nines availability, and you don’t want to put in boatloads of redundancy into the network – because of course that becomes counterintuitive or economically prohibitive – then that’s where the role of AI and ML [comes in].

Especially on the edge – some of those edge applications will be running real time, whether it be on the network side within the radio interface, or whether it be some interface between the edge and a local application within a factory or distribution environment. It’s going to be critical, a huge role in the future.

You mentioned the importance of ethically deploying AI. How much does IBM and firms like it think about where and how it could all go wrong?

Well, we spend a lot of time thinking about it. In fact, if you look at the IBM website now you would see a whole campaign that we’re driving at the industry level – at the cross industry level for that matter –  on the ethical use of AI. It is incredibly important, where does it really impact the telco sector? It’s really in the design process. If you are running a service which has universal access, the service was designed effectively to the lowest common denominator, but as we go towards more specific use cases that have robotic systems and humans interfacing with each other, then part of the design criteria has to have AI front and centre to that.

And then you have to model out where the things could potentially go wrong, in time and motion studies for example. So I think the big differences is in new design processes and new innovation processes. You will actually see constellations of expertise coming in, so you’ll see finance people and lawyers and you’ll see designers and creative marketing people and they all need to come in because they all see the world through those particular evaluative perspectives, and I think when you have that in, that’s how you pick up the issues of where things can go wrong. You’ll intersect them early.

Unless you are completely involved in that world, it’s sometimes hard to track where we are with advancements in AI because it’s been surrounded by so much hyperbole for years. Could you describe what has changed significantly in the last five years, and in another ten year’s time where will we be in terms of AI integration into society?

If you look at one of our products, Maximo Visual Inspection for example, it is able to pick up on physical anomalies in ways that you couldn’t possibly pick up before, because either the distribution line or the or the production line was running too fast, or the human eye just couldn’t pick up the types of microscopic issues and whatever else.

Those types of things are truly advanced and I think we should be very excited. Are we going to see a world of robots where suddenly all humans become useless? No, I don’t think so at all. And in fact, if you look at in history since even the first industrial revolution, it was quite clear that those that were impacted were pretty upset but actually in the long run, history tells us that every time there has been some level of an industrial revolution, actually, humanity has become better. Jobs have actually increased as a result of it.

The question is what type of jobs exist in the future, not whether there are less jobs or not. But again, I think it also goes back to a balance between corporate social responsibility, responsibility to things that matter to most of society and the ability to generate profit. In fact, I’m much more encouraged that we’ve become far more conscious of that consideration than we ever have in the past.

You’ve also partnered with Walmart to use blockchain to track supply chains recently. What other real world scenarios could blockchain and AI be mapped onto in the future?

One of the most exciting things is not just where you’re seeing blockchain or web3 type technologies come in, but actually where you’ll see them coming at the intersection of private and public networks. So if you think about it, today most distribution systems are within contained environments. And all the approvals of that, whether it be commercial approvals, or whether it be operational approvals – or denials for that matter – they need to happen in a way that requires a lot of human intervention.

If you look at some of the Bell Labs studies for example, you can see that there is still a ton of industry that is physical, that actually still work in highly, highly manual ways of working. So when it comes to infusing these technologies, if you can imagine that what we will have is smaller environments and let’s say, local domains.

Let’s imagine you went to Saskatchewan up in northern Canada. You’ve got a mining town up there, and most of the town operates in and around that environment. There will be private networks within those domains, there will be private network in the mine, there will be a private network in the supply chain, there will be a private network in the emergency services, and at some point or another, you will be able to tie those networks together through a data fabric and then you will actually have blockchain running through that data fabric.

We’re right in the middle of enabling all of that. The beauty of that is then that the existing manual interdependency between industry sectors will actually start to become much more operationalized end to end. And so if you’ve got emergency intervention required, or you’ve got IoT devices that are sensing in the mine and something goes wrong, you don’t necessarily need the mine person to tell you that there’s something wrong because the emergency services will have interconnected data fabrics which will allow them to intervene way before maybe even the people in the mine know about it. So I think that’s where I’m very excited, about how these technologies start to infuse and come together. Again, the critical thing will be trusted AI environments, trusted data environments, and a lot of security.

Check out the other half of our interview with IBM, in which Rose says there are not many firms out there that understand the telco sector.

Get the latest news straight to your inbox. Register for the Telecoms.com newsletter here.