Hack the Plant Episode 14: AI and Critical Infrastructure
When will hard infrastructure have machine learning capabilities? It might be sooner than you think.
In this episode of Hack the Plant, I am joined by Ariel Stern, formerly an engineer in the Israeli Ministry of Defense and a civil infrastructure project manager, currently CEO of Ayyeka, which offers remote monitoring for industrial Internet of Things systems.
Ariel has a forward-looking approach to creating resilience in critical infrastructure…anticipating that we are entering a new era for critical infrastructure….from IoT data creation, management, and analysis to advanced Artificial Intelligence pattern recognition and prediction.
Is this science fiction? Join us to learn how the technology that can create resilient infrastructure for tomorrow is here – today.
(Subscribe to Hack the Plant on Spotify or Apple, by RSS feed or search for it wherever you listen to podcasts.)
Joshua Corman:
Our dependence on connected technology is growing faster than our ability to secure it, especially in areas affecting public safety and human life.
Bryson Bort:
I’m Bryson Bort. And this is Hack the Plant. Electricity, finance, transportation, our water supply. We take these critical infrastructure systems for granted, but they’re all becoming increasingly dependent on the internet to function. Every day I ask and look for answers to the questions. Does our connectivity leave us more vulnerable to attacks by our enemies? I’m a senior fellow at the R street Institute and the co-founder of the nonprofit ICS Village, educating people on critical infrastructure security with hands-on examples, not just nerd stuff. I founded GRIMM in 2013, a consultancy that works the front lines of these problems every day for clients all over the world.
For today’s episode, I’m joined by Ariel Stern, formerly an engineer in the Israeli Ministry of Defense and a civil infrastructure project manager, currently CEO of Ayyeka, which offers remote monitoring for industrial Internet of Things systems.
Ariel has a forward-looking approach to creating resilience in critical infrastructure…anticipating that we are entering a new era, where hard infrastructure has machine learning capabilities.
Ariel Stern:
We are in the beginning of a new revolution…..now everything began to be interconnected and I think this is the big [inaudible 00:01:28] revolution. About interconnecting, all the different assets out there to one big network for decision-making for better operation.”
Bryson Bort:
It might sound like science fiction, but the technology that can create resilient infrastructure for tomorrow is here today….from IoT data creation, management, and analysis to advanced Artificial Intelligence pattern recognition and prediction.
Ariel Stern:
The infrastructure space, it is majority is still in the stage of creating the data. Meaning you need to have a systematic manner to collect the data or create and collected the data, and then manage it before a big deployment of AI tools and machine learning tools are utilized and we are not there yet in the majority of the utilities. There are some outliers out there that are dramatically better than the others, but the majority of the utilities in the remote assets can dramatically improve the way they collect and manage data.
So the algorithm stage is still in its infancy, meaning we are not at the very big and complicated area yet in this space. Probably five to 10 years into the future is the bigger part around those areas.
Bryson Bort:
How can technology make our critical infrastructure more resilient? How does AI fit in? We explore this, and more, in this episode.
Bryson Bort :
Industrial revolutions. In the beginning, we discovered mechanization through steam and water power, then mass production, assembly lines, the introduction of electricity, finally, industry 3.0, automation, computerization. A lot of the challenges and things that we’ve been talking about on this podcast over time but are we on the cusp of another revolution? Ariel, what do you think?
Ariel Stern:
Yes, we are in the beginning of a new revolution. The 4.0 revolution, the fourth industrial revolution that affects everything. As you mentioned, the three previous revolution brought many things to live, but now everything began to be interconnected and I think this is the big Internet of Things revolution. About interconnecting, all the different assets out there to one big network for decision-making for better operation. This is a thing, the meaning of the fourth revolution.
Bryson Bort:
But it goes beyond interconnection. We’re looking at where our devices can start to become autonomous, where industrial cyber can begin to make its own decisions and react to what we see in the environment?
Ariel Stern:
Yes. So interconnection is the first part and by interconnecting, you have the ability to make better decision through awareness. The devices or the assets start to understand or understanding is… There is a big difference between machines and humanity but almost understand what is going on and you can program them or define set of rules… Sometimes it’ll be better reactive to the environment. So yes, interconnections the beginning, then you get into much more interoperable and awareness mode.
Bryson Bort:
Okay. So it sounds like we’re also now starting to look at the concepts of machine learning and artificial intelligence. Can you explain what those are and why are we doing this and how does it benefit us?
Ariel Stern:
So, yes. Machine learning and artificial intelligence are big terms and it will be very hard to isolate one sentence. But in general, when you have a lot of data and you have a lot of need to use this data, in many cases, running standard processes or standard algorithms is not enough. You need to go to the higher level of algorithms and develop the system to represent what we call intelligence, meaning their ability to deduct meaning from those patterns of data to look for things you cannot see with your eyes, meaning interconnection between data.
Ariel Stern:
It’s strange, but many of us are using artificial intelligence tools on a day to day, without even knowing that. For example, the camera application in smartphones improves video or photo quality for artificial intelligence. The search mechanisms on the web. When you look for something and the search engine improves your searches, all of those are examples of a day-to-day AI, artificial intelligence, which help you to do more with the data or the processes around you. Same goes for infrastructure. The ability to do more with the data collected and to react better to changing conditions, this is the essence of the AI.
Bryson Bort:
So what brought you into this? What is your background?
Ariel Stern :
So I started an early career in the defense industry and multidisciplinary project management, and then spent the middle part of my career in civil infrastructure. I was a project manager for a very big transportation project, where I saw the different inability of different systems to work together. And after a few years in project management in infrastructure, we started Ayekka. My friend and I started Ayekka. He comes from the cybersecurity and software world. I came from the hardware and project manager and we started Ayekka through that.
Bryson Bort:
So you were a project manager and industrial control systems, that was what got you into infrastructure.
Ariel Stern:
Actually, it was more than that. I was a project manager in a civil engineering project, although my education is electrical engineering. I did some project management for civil infrastructure. So I had the opportunity to see the multidisciplinary world of both civil infrastructure and design of electrical systems, and then connect the two work together.
Bryson Bort:
Okay. And so at what point did you realize that we were on the cusp of this fourth industrial revolution?
Ariel Stern:
I don’t think that you can say it was a single point, but it’s through the process of understanding the landscape and getting to know who is doing what. We developed the concept that things can be done better and bringing concepts and disciplines from the defense industry into the civil infrastructure space brought the idea that many more benefits can come out of embedding intelligence and data from remote field assets into the day-to-day operation of infrastructure. So it was not a single point in time, but more likely a process during the stages of Ayekka that brought us to this direction.
Bryson Bort:
So in terms of that timeframe, what’s going to be coming next?
Ariel Stern:
Well, it’s a big question. I’ll tell you what we see. Our solution started as a very efficient way to create data. This is the first challenge we faced. How to effectively create data [inaudible 00:06:38]. And then, the second part of the solution came to managing the data. How to make sure the data is stored properly, protected and interoperable with other data systems.
And I think that now over the last few months, we are starting to face the utilization stage, meaning now we have a huge repository of infrastructure data, how can we bring more benefits to the customer, which is more than the obvious data tools or set of tools coming from the data directly?
So we started to build a set of solutions around two themes. The first theme is anomaly detection, meaning to detect an abnormal condition in the infrastructure before you see damage or any of the obvious evidence. And the second area is about data quality, because when you say the fourth industrial revolution, it’s around data and you need high quality data and we have years of experience for creation of data and improving the data quality is a very, very big challenge in the space. We are also focusing on this area, data quality and anomaly detection.
Bryson Bort:
One of the things that’s unique about critical infrastructure is the fact that we’re not just talking about computer data in terms of what we typically see in computing, right? It’s not just my phone storing something. It’s not my laptop just having something. It’s computing that can actually change the physical world, hopefully most of the time for better, but as you said, anomalies can be where something bad might happen. We’re talking about the potential of loss of life or limb. So can you translate and give us more examples of what you mean by data in that context?
Ariel Stern:
Yes, and before we dive to more specific examples, I would like you to imagine how infrastructure looks like. Because for many people it’s not so obvious when you talk about critical infrastructure or hard infrastructure. So we mentioned the water distribution networks or the wastewater collection networks or the energy networks. We’re talking about very, very big dispersed networks and effectively collecting data from them can bring a lot of benefits on the day-to-day operation.
We have, for example, very interesting and unique projects around stormwater management, meaning the real-time management of rain in big catchment basins, and how to prevent or reduce the number of overflows and floods in specific metropolitan areas. This is one great example for how smart infrastructure can really save lives and save money.
And the second one I can give is for example, water quality. Monitoring in real time the water quality in different parts of the network, meaning replacing manual processes and detecting the degradation or deterioration of drinking water quality in real time. This from the water space and on and on. Every type of infrastructure have its own unique set of challenges. They share the same common challenges, but they translate into unique sets and collecting the data and using it for algorithms and software system is dramatically improved the way the stakeholders managing the infrastructure.
Hopefully it gets some color to add to the story because I understand that connecting the very physical infrastructure… In your mind, you probably see in your mind manholes and power lines and taps and roads and bridges. This is the image people think about when you talk about hard infrastructure and now try to imagine what is the digital space looks like. So you think about computers and computer screens and databases and putting those two together is sometime mind-blowing. Connecting those two physical and virtual work together. But this is what we do on a day to day and it really changed the way operators and stakeholders are making decision over utilities.
Bryson Bort
So going back to where you talked about data quality, what is data quality?
Ariel Stern:
Let’s look for a specific example. Let’s say you want to measure… Let’s go for the very timely subject of floods. Right now, everywhere in the world, we are in August 2021… Everywhere in the world, you have major floods. So running a better collection system involving monitoring level and flow in wastewater collection networks.
Ariel Stern:
So when we talk about data quality, we say those remote monitoring stations that measure those physical parameters, both of the level and the flow and other parameters of the network, you need to understand that in many cases, the data is not good. The data itself cannot be declared as good data because it’s inaccurate, maybe not available. Maybe you have some technical problem. Maybe the specific site is not very tuned to provide high quality data and when you take this data and utilize it into artificial intelligence system or decision-making protocols, you need to make sure that the data you can claim is good quality. Exactly like the sound quality, video quality, any type of other digital data that you collect, you will need to monitor its quality, same with critical infrastructure. But unlike more common thing that we are dealing with, meaning in the IT space, in infrastructure, all of those quality issues are becoming 10 times harder because everything is dispersed.
Bryson Bort:
Okay, Ariel, this sounds good. We’ve got computerized systems. This per se produces efficiency. This provides better insights to make these decisions, but it also introduces risk. How do we look at the security aspect of this to reduce the greater risk that we’re bringing to interconnected devices?
Ariel Stern:
You got it correctly. Meaning connecting things. Also, bring the concern of cybersecurity, which is a very big term. In critical infrastructure, if you have cyber threats, they can create a lot of damage over very, very short attack time stents. They can shut off critical services. They can create havoc in the networks. They can disable operator ability to access their facilities, et cetera.
We are facing now with the infrastructure space, they are battling at two fronts. One front is the operation side. They are battling daily with stress networks, with compliance with the operation side of the network. On the second hand, they also battling on the cyber front against endless and countless attacks from different entities. And we need to make sure that when we connect the assets, we make sure that we take care of the cybersecurity aspects, the same as we connect the laptop or computer inside the facility network.
And this is an area that some people are currently not focused on. Meaning the IT department is very good at connecting new network assets. For example, new laptops or servers, et cetera. They have protocols and many, many routines around how to connect new computers or smart phones, those things. And when you connect remote asset infrastructure assets, the manhole or the emergency generator or the storage tanks into the network, and you literally embed them with connectivity and intelligence as part of the revolution, you need to make sure that those intelligent assets and connected assets are well-protected at least as your laptops and smartphones and servers.
We are currently at the frontline of explaining and educating the market that when you embed in remote physical assets, typically there are hardware only, what used to be only hardware with electronic connected devices make cybersecurity one of your top priorities. On top of almost everything else… Because those assets become critical part of your IT network as well.
So you’ve got the story very right, meaning the concerned about cybersecurity in critical infrastructure should be a first degree priority for decision makers and operators.
Bryson Bort:
So artificial intelligence and machine learning… Fact, we kicked off the podcast series with P.W. Singer who wrote a book called ‘Burn-In’, talking about how the abuse of artificial intelligence led to dystopic results. How would you address that perspective as we start to include artificial intelligence and machine learning and the fourth industrial revolution?
Ariel Stern:
I think that for those who are involved in this process, not to be very careful about how they embed such tools into their day-to-day routines. Artificial intelligence and machine learning are description of tools or concepts. There are many types of specific tools inside it. So you need to be very careful how to do it in order not to lose control of the network and not to lose the sense of human intelligence inside the framework of infrastructure, but take the benefits of AI machine learning and properly embed them into the decision-making process, but in a controlled manner. Meaning, some people think that the infrastructure space is moving slower than others.
There is a lot of saying, why aren’t they adopt new technologies? How come they are still in the middle ages, et cetera? But the reality is that in many cases, they are dealing with processes and networks that are much more important than others. That embed technology very, very fast or make a huge experiments through AI and machine learning.
In the critical infrastructure, safety is the first priority and then security is the second priority. You need to be very careful how to embed the tools in order to make sure you get the benefits but reduce the risk. Giving an example in our case. Improving data quality does not increase the risk because it’s almost a passive process. You improve data quality and then you use it further down the road. Making sure that AI and machine learning are well controlled, well-trained and well managed is another challenge that need to be properly handled by the decision-makers in the infrastructure space.
Bryson Bort:
One of the challenges with machine learning and artificial intelligence is the quality of the data, but in a different way. There are things that we don’t know, and that creates implied bias in our design. We have to train the computer to think in a certain way and those data sets help do that. Does that play at all into how you work with data and helping provide training sets for that?
Ariel Stern:
Yes. We started to see that, but from our perspective, those challenges will be down the road. We are still much early on in the process of taking infrastructure data and utilizing it through machine learning.
The infrastructure space, it is majority is still in the stage of creating the data. Meaning you need to have a systematic manner to collect the data or create and collected the data, and then manage it before a big deployment of AI tools and machine learning tools are utilized and we are not there yet in the majority of the utilities. There are some outliers out there that are dramatically better than the others, but the majority of the utilities in the remote assets can dramatically improve the way they collect and manage data.
So the algorithm stage is still in its infancy, meaning we are not at the very big and complicated area yet in this space. Probably five to 10 years into the future is the bigger part around those areas.
Bryson Bort:
Okay. So, we’re not quite there yet. Is there a particular industry that you think is going to adopt this more quickly than others?
Ariel Stern:
It’s come in different directions. Meaning, every industry has its own reasons why they’re adopting it. Some of them are going for the financial benefits. They can save money or dramatically optimize. Some of them are going for the environmental aspect because they are challenged by the environment and some of them are only facing it when there are compliance related issues. Meaning when the regulators are demanding it.
We see a great traction in the water and wastewater space on a global scale, meaning north America, far East Asia, et cetera. We see some traction in the oil and gas in the upstream oil and gas, but I think water and wastewater is the biggest growth engine of this space for the upcoming three to five years. Gut feeling.
Bryson Bort:
And why is wastewater going to be what goes first?
Ariel Stern:
I think that the water and wastewater segments are facing unparalleled challenges from different fronts, such as operational stresses, meaning you need to supply much more services to your population. You have environmental aspects, the global weather pattern changes. You see more rain, in many cases, floods. You see less rain, in many cases droughts further down in the Southwest and also compliance, meaning regulators are very involved in the water and wastewater.
So I think this push them to the reality that they need to invest in solutions that are not just replacing the infrastructure or rebuild it in a bigger way, but in adding intelligence to it. So our belief from our experience, those are the segments that would benefit most immediately from such a transition.
Bryson Bort:
Okay. So it’s going to be a not internet connected crystal ball. It has magical powers. It has all of the future data sets for artificial intelligence in it to predict. What do you think is one good thing and one bad thing that’s going to happen in the next five years in critical infrastructure?
Ariel Stern:
The good thing is that I think now it’s a top priority for policy makers. Meaning, everybody understand, and they’re deploying resources to improve, rebuild and reinvent infrastructure. It started with the new deal that is now taking place in the US but it’s all over the place. So this is the good thing. Infrastructure become a first priority.
The bad thing is that I believe that some individuals are underestimate the need to dramatically change the way infrastructure work. Meaning they believe that yes, infrastructure is a priority, but you need to rebuild it and not to dramatically change the way it works. As long as this type of mentality will be popular in infrastructure space, think we’ll not move to the direction that will change the infrastructure.
So the good thing is that there is no focus. The bad thing is some people or some individuals or some entities still need to rethink how they address those challenges in order to make the complete transition into the benefit of digital space.
Bryson Bort:
All right. So you can be king for a day. We don’t even need to worry about the policymakers. Now you have an internet connected air-gapped, if you will, magic wand. Wave that magic wand. What’s one thing that you would instantly change in this industry?
Ariel Stern:
Probably the requirement that a certain percentage of the infrastructure assets under a certain management entity will be interconnected within certain period of time. So to mandate the transition to digital space by addressing the unconnected part of the network, meaning telling your utility, you have, let’s say 36 months to connect 20% of your remote assets. Things like that. I think this will dramatically kickstart a very, very massive flywheel for digitization.
Bryson Bort:
All right. So what’s something that we didn’t cover that you would like to talk about?
Ariel Stern:
I think it’s important to emphasize that when you talk about industrial revolutions, the key word is the industry and infrastructure is a bit different. And it’s now the rise of the term infrastructure 4.0, which is a bit off shot or a spinoff of industry 4.0 is understanding that the infrastructure needs special treatment or special care in how they deal with the basic part of IoT, meaning the interconnectivity, the connection of the assets, the management of the data, the protection from cyber security, the disconnection from the rest of the world. All of those aspects are quite unique for infrastructure.
And it’s important to emphasize that yes, IoT is moving forward. Yes, we see a lot of industry use cases and we are only starting to see the full potential of infrastructure IoT, meaning IOT 4.0. The more focus we will be put on how to connect those remote infrastructure and how to better utilize algorithms and methodologies to better manage it, I think we will move faster toward a more sustainable infrastructure. Because the way things are managed today, a lot of people believe this is a dead end, meaning there is not enough capacity in the existing networks to sustain us for the upcoming three, four, maybe five decades.
A shift should start now and digital infrastructure is a major component in this shift.
