 Good afternoon, everyone. I see it's 1.30, so we are going to get started. It is my pleasure to introduce our chair for this session, Human in the Loop, the Changing Role of Humans in New and Advanced Reactor Designs. Our chair is Dr. David Desaniers, who is a senior level scientist at the United States Nuclear Regulatory Commission with over 34 years of experience in the nuclear industry. For the past 15 years, he has served as the NRC's Senior Technical Advisor for Human Factors and Human Performance Evaluation. His contributions include the development of guidance, standards, and regulations that address the nexus between the design and operation of nuclear power plants, human and organizational performance, and the protection of public health and safety. He has held leadership positions within the Institute of Electrical and Electronics Engineers Nuclear Power Engineering Committee, the American Nuclear Society's Large Lightwater Reactor Consensus Committee, and the Nuclear Energy Agency's Working Group on Human and Organizational Factors. For his achievements in advancing nuclear safety and innovation, Dr. Desaniers is a recipient of the NRC's Distinguished Service Award and the American Nuclear Society's Don Miller Award. Please welcome Dr. David Desaniers. Thank you, Stephanie, and good afternoon, everyone. I hope you had an enjoyable lunch, perhaps catching up with an old colleague or perhaps making a new acquaintance. If it was a new acquaintance, I wish that person many years from now is one of your old friends. The RIC is a wonderful opportunity to meet people who can enrich our professional lives and share information that help us achieve our goals. I hope you are taking full advantage of these opportunities. About a week ago, I looked at the registration for this session and was elated and frankly a bit daunted to see that we had over 700 attendees registered for this session, either online or here in the room. And I had to sit down for a minute and do the math, and it dawned on me this audience is about 700 times larger than what I'm used to speaking to. The engineers in the audience have probably done the reverse calculation and figured out that means he's used to speaking one-on-one with people, which is true. That's my preference. So if you'll indulge me, I'm just going to imagine that this is a one-on-one conversation with you all. That is 700 of you in parallel. And feel free to do the same if it makes you feel more comfortable. It is my honor and pleasure to chair this session of the RIC. It's an honor because I get to share the stage with a wonderful panel of speakers. They are bringing to our discussions today a tremendous depth and breadth of experience in the application of human factors to the nuclear power domain. It is my pleasure to chair this session as they will be discussing a topic that I believe is exciting, relevant, and important. Those of you who have heard me speak in the past have noticed that rather than quickly moving past the title slide, I like to dwell on the words of the title as a means of introducing a topic. This session is titled, Human in the Loop, The Changing Role of Humans in New and Advanced Reactors. Those of you who have been around a while, either in the field of human factors or perhaps design testing, are likely familiar with the older term, man in the loop. Now, we did not use that term in part because quite some time ago it was recognized man in the loop as a sexist phrase. More importantly, human is a term that is much better at capturing the rich and diverse considerations that should be taken into account when considering the role of individuals and groups of individuals in the operation of nuclear power plants. So why do I say that? For many of you, the phrase man in the loop is likely to conjure up the rather dry list comparing the capabilities of men to the capabilities of machines. The acronym was MOBA MOBA. Men are better at, machines are better at, hence MOBA MOBA. In these MOBA MOBA lists, you would likely find items such as machines are better at repetitive tasks. Not surprising, machines don't get bored and their attention doesn't fade. Another example, machines are better at doing complex calculations. But now I want you to consider what might seem to be a philosophical question. What does it mean to be human? What makes you human? Each of you may have a somewhat different answer to that question. Whatever the answer, I expect if you now think in terms of human in the loop rather than man in the loop, your thinking expands beyond basic considerations such as whether a man or machine is better at doing repetitive tasks. With that, I hope I have wetted your appetite for our panelists' views on how the role of humans is likely to change in the operation of new and advanced reactors and what might be some of the most important considerations or implications of such changes. Now, before we get to those presentations, though, I want to take a moment to acknowledge the individuals that have made this session possible. Those are our session coordinators, Stephanie and Nea. Would you please stand up? The coordinators for this session, these are my wonderful colleagues, Dr. Nea of Hughes Green and Dr. Stephanie Morrow. Both are accomplished human factors psychologists and researchers in NRC's Office of Nuclear Regulatory Research. You've already given them applause, but let's give them another applause for bringing together this wonderful session. There is a lot of work that goes on to make these things possible, and I really appreciate all the work they put into it. Now I'm going to briefly turn over the mic to Stephanie, who is going to speak to us a little bit about QR codes and other logistics. Sure. Can we have the QR code back up on the screen, please? And for this session, we encourage you to submit questions throughout the panel. For those in the session room, please take a moment to scan the QR code displayed on the screen so that you can participate in both the Q&A and the live polling during this session. Again, take a moment right now to scan the QR code. It'll be much easier as you go along. For those of you joining virtually, you'll find there's a box to the right of your screen with two tabs for electronic Q&A, live polling, and feedback. You can direct questions to either a specific panelist or what we would ask is there is an option that says USNRC staff. If you have a question that would apply to any panelist, please select that option, USNRC staff, to pose a question to the entire panel. We will also be incorporating live polling questions into our session today. For those of you in the room, you'll find a poll option on the same web page where you submit questions on your mobile device accessed with, again, the QR code displayed behind me. In-person participants can also participate by texting their choices to 22333. When we introduce a polling question, there will be a slide displayed in the room with the instructions for participants to text their responses. For those of you participating virtually, once you've logged into the virtual conference platform and entered into the session, there, again, is a tab to the right of your screen for live polling questions. The polling results will be displayed on the screen to reflect responses from both our in-person and virtual session attendees. Thanks, Dave. Thank you, Stephanie. Okay, hopefully you've all had your opportunity there to get the QR code, so I'm now going to move on to the next slide. Just going to, at this moment, provide a brief overview of our panelists. I'll provide a more fulsome introduction of each of them as they come to take the podium. What I want to point out at this point is that, as I noted before, we have a great depth of experience and diversity in our panel here. Kicking things off for us will be Karen Priestman of the Canada-based nuclear energy engineering service firm Nuclear Promise X. She will be followed by Richard Pace of the power and energy consulting firm Sargent & Lundy. Then we'll have Claire Blackett currently of the Risk Management Consulting Group Risk Pilot and formerly a researcher for the Institute for Energy Technology in Norway. And Claire will be followed by Paulo Savioia of the Stook Radiation and Nuclear Safety Authority of Finland. I will provide some remarks to follow. But again, collectively our panelists will bring perspectives from their work in different types of organizations as well as different countries. Each will make a brief presentation and then we will make plenty of time available for audience questions. So those presentations that you will hear from each of our panelists, we've asked them to adjust this particular question as part of their presentation. What is the greatest human factors challenge or opportunity we will face in the nuclear industry in the future? Before we begin those presentations, we have a polling question for you, our audience members. The question will help our panelists understand your general familiarity with human factors, a field of study and a discipline that is central to the topic of this session. So could we put up the first polling question, please? So your question is, how familiar are you with the field of human factors? Glad to see that we have at least some understanding, but not surprised to see if we have, in some cases, very little understanding because it's been my experience over many years that there's many people out there that this is kind of a foreign discipline to them. But as these results continue to come in, it is now my pleasure to introduce our first speaker, Karen Prizman. Karen started her career in Ontario Hydro, the precursor to Ontario Power Generation, and the company currently building North America's first small modular reactor. Throughout her 33-year career, Karen has worked and still going, right? 33 and still going. Has worked in a wide variety of roles for consultancies, design organizations, and manufacturers, and for 10 years as an independent contractor. Karen's career has been a mix of both systems and human factors engineering, largely in high hazard industries, including defense, rail, and oil and gas. For the past 18 years, Karen has worked in the civilian nuclear industry, most recently for Japan-based Hitachi GE and US-based GE Hitachi, for whom she led high-profile and extensive human factors engineering programs as part of their new build projects. In January this year, Karen joined Nuclear Promise X, a company focused on supporting implementation of new digital technologies within existing and new plant designs. Karen has a Bachelor of Applied Science and Systems Design Engineering with a human factors minor from the University of Waterloo in Ontario, Canada. She is an Ontario licensed professional engineer, a Chartered Fellow of the United Kingdoms, Chartered Institute of Ergonomics and Human Factors, and a member of the International Council on Systems Engineering. Please welcome Karen Prizman. Thank you, Dave. And thank you everyone for attending this session. We know you have a choice of sessions to go to, so we appreciate you coming to this one. Dave's comment about speaking one-to-one, that light is so bright. It pretty much limits who I can see, so it is almost like I'm speaking one-to-one. If we all have kind of a bit of a deer in the headlights, look, that's why. We've also got, just for fun and games, we've got a timer there, so talk about time pressure. But I will go fairly quickly through my slides. We want to leave enough time. I've just reminded Nea to start it. Through my slides, just a bit about Nuclear Promise X, because we are a new and smaller company. And then I will address the question that was up on the slides, Dave's slides. Let's see if this works. Here we go. Nuclear Promise X was formed about five years ago, and basically found it on the firm belief even five years ago before this renaissance that we're going through, that Nuclear was essential for the global climate emergency and for the goals of 2050 of reducing emissions. And in order for that to happen, it is also our vision that all that we do in Nuclear has to be done better, but also faster and cheaper. And I think that ties in with a lot of the content of this RIC. So we're pleased that that fits really well with us as a company. We started with, and I'll just move quickly through these, we started with sort of a products area, a lot of digital transformation, a lot of robotics and digital twinning, and even processes, making highly digitalized processes that were quite document-centric before. And recently, the last two years, we've delved into the services, we do project management, using our digital tools, of course, and design engineering, which is now my baby. And my goal is to move the needle on how we do design engineering in the nuclear industry. So I have a stated agenda. I'll just put that out there. And I love this conference for that, you know, looking forward. So I will skip all those. And we do, you know, we do a lot of apps and software and digital engineering. We have NPX AI, which helps to even, you know, engineering processes that are very data-centric. You know, we can use AI in order to make those processes more intelligible and faster and efficient for people. So, down to the meat of the matter. HF, and I want to preface this by saying that these are both challenges. I've put them as challenges. They're challenges and opportunities. They really are. You know, yes, they're challenges because we haven't traditionally looked at these areas, but they're opportunities because we can now focus on these areas. We have to focus on these areas. If our new builds are going to be run safely by the people who are running, you know, these plants, then we've got to consider these things. And the first one is, at this point, there's a lack of driver's tools and skills needed for not just doing HFE in the modern digital design world, but integrating effectively, integrating early. We have struggled with that as a profession from the very beginning. It's going to get even worse because it's going to go faster. Decisions are going to be made up front that are going to have a huge impact. So, and I think that the tools we have as human factors practitioners aren't necessarily the tools we need. There are a lot more nuances in digital, you know, human AI teaming, human robotics teaming and just even just digital interfaces. The traditional rules of human factors just can't be applied. And the analysis we do is different too. Workload and staffing is a whole different ballgame when you're talking about highly digital plans. The challenge number two, along with the tools and the skills which we're missing or that we need to get better at, I don't think we have given enough thought. Generally, people tend not to really understand or be able to quantify or discuss human factors risks, especially future. We're pretty good at it in human factors practitioners, although even then not across the board, I think that's just going to be a whole other set of risks, different risks. They don't, they're not removed by removing some of the human in the loop. They're just moved to other places in the loop. We always talk about that. You don't get rid of the human, they're just somewhere else. And so I think understanding the nature of the risk is going to be important for where we focus our human factors expertise and effort because number three, and there's some more things on the slide, some specific examples of where we've underestimated the risks. And I think there is a tendency we're going to overestimate in some areas of traditional atrophy, we're going to underestimate in a lot of areas. And then I think the third one is, and I had something on here too about the people that we have and having enough people. And move where atrophy is a profession and start to use some of the tools and the digital aids that we have that we're putting into the plants, we need to start using them within not only the people doing atrophy in the design and new build, but within the regulators. We need to start using these digital tools, but then also understanding the effect on the people in those organizations and what those have. But we need to start leveraging things that other disciplines, particularly systems, and yes, I am both systems engineering and human factors, but I know that they too are not separate, and there are tools that are used in systems engineering that could really help human factors. They shouldn't be separate anymore. So that's my other thing. I think that the one I didn't really touch on which was part of the second slide, which I'd like to say, it's the rate of change that's the issue. I know that all the regulators are looking at, yes, we have to modernize, we have to look at human factors differently, there's going to be AI, but that's only going to get quicker, and I don't think we're responsive and quickly enough right now. So we need to start using digital tools to be able to pivot quicker and be more agile in how we do human factors in the industry. And that's it. That's all I do for time. Thank you. Okay, and if I didn't make it clear up front, I don't remember what I said two minutes ago now, we'll take questions at the end of all the panelists speaking. Our next speaker is Rick Pace. Rick has 13 years of nuclear experience and three years of human space flight operations experience. Rick is currently the lead human factors engineering consultant within the Sargent and Lundy Digital Modernization and Consulting Group. In this role, Rick is setting industry, technical and regulatory precedents, working with multiple utilities on all current major control room design modernization projects. Prior to Sargent and Lundy, Rick was the lead instrumentation and control human factors licensing engineer for the Westinghouse AP1000. And the overall technical licensing lead for the Westinghouse EVINCI microreactor project. Rick is a member of the Institute of Electrical and Electronics Engineers Working Group 5.2 on human factors engineering and the Nuclear Energy Institute's Digital Instrumentation Control Regulatory Task Force. He holds a Bachelor of Science Degree in Aerospace Engineering and a Master's of Science Degree in Physics. Please welcome Rick Pace. Thank you for having me. It's an honor to be here today. So today we are talking about the changing roles of humans in new and advanced reactor designs. And so we're asked to answer the question, what's the biggest challenge in this? And I think there are many. But if I could name one, I think one of the biggest challenges, also being one of our biggest opportunities, is the increased use of automation within our systems, within our plants. Because as we increase the use of automation in our plants, this obviously is going to have a direct impact on how the human and how the operator interfaces with this automation and interfaces with the system. And as I pondered this question, it's not that some automation and using automation is new, it's just the increased level of our ability now with new technology to leverage this automation to its fullest and to a greater extent. And so that's what I'm going to talk about today and so I'm going to use this my time here just to give a few thoughts on things that we need to consider as an industry as we move forward to increasing automation as we look forward to the future. So the first thing I'll say for the increased use of automation is the opportunity to improve the safety of our plants. So one of the biggest contributors to system and plant heirs are human heirs and that's understandable. Humans are fallible and we make mistakes. And so if you can minimize the amount of times that a human or an operator touches the system, you could theoretically minimize the amount of heirs that get introduced. And so as we rely upon automation, we can reduce the amount of times that the operator or the human needs to interface with the system and so that could be a positive in plant safety. Now on the other side, the more and more automation that we use could introduce more and more complex systems which in itself could also introduce system heirs, right? And so there's a balance that we have to make in my opinion on using the right level of automation for a given system. And that's an important point I think that I would like to stress is, you know, automation isn't an either or thing, right? So it's not either you're using it or you're not using it. There's, you know, there's gradations of the amount of automation that we could use in our plant. So at the extreme end, you could think of an automation level of one being fully manual and the operator does everything that the system and the plant needs or an automation level of ten say where the system does everything and completely ignores the human in the loop. Now in reality, we're somewhere in between that so you can consider or think of, you know, varying degrees of automation within those two extremes such as automation with operator supervision or operator consent or maybe the system performs its automation and it decides when the human should be in the loop and when the human needs to be informed of something or the human needs to take action. And so you can sit here and think through and others have the different levels of automation. I think the point here that I want to make is that, you know, as we leverage this tool, we need to think of what makes sense for the given situation and apply that and understand the impacts because ultimately at the end of the day, no matter how much automation we use and leverage, you know, the operator still needs to be aware of what the system is doing. They need to be knowledgeable of the system, their situation awareness is one of the challenges that we have to overcome with this because, you know, as the system is performing tasks in the background, we always have to make sure that the operator understands what's happening with the plant, right? They need to be able to troubleshoot any failures and ultimately if automation fails, they need to be trained in such a way where they can take the appropriate action to keep the plant in a safe state. And so that's another sort of sub-theme that I'd like to point out is training is an important piece to this. You know, the rules of the operator are going to change as we use more automation, which means that their training and their understanding of the systems are going to have to keep up with that. The last thing I'll mention here is a very positive thing for automation is improved fault detection. So self-diagnostics in these systems can continuously monitor the system and ensure that there are no faults. And, you know, Dave had mentioned it earlier in his speech. But the system can do this more effectively besides relying solely on human intervention, right? The system can sometimes see things that the operator can't see even before they become a problem to the plant. And so there's a ways to leverage this so we can have early fault detection so that we can take action and keep the plant in a safe manner. Again, the operator needs to be trained on these things and to be able to troubleshoot and diagnose when these self-diagnostics aren't working, but we can leverage those for improved plant safety. So ultimately, we need to think outside the box and leverage this technology. You know, I've seen firsthand where we spend a bunch of effort in time modernizing our systems, yet we don't take full advantage of the benefits of the system that we put, that we install. And so self-diagnostics is one of those areas. And I'll just end my talk here. Encourage us as an industry to look forward to other industries because, you know, other industries are doing leveraging automation to a fuller extent than we are. And they're designing these remote systems, remote operating systems already. And I know that that's the direction that some of us are going to in the human factors world is remote operations. And so I think that's important for us to leverage those lessons learned. Dave had mentioned that I did have some time and experience as an operator for the International Space Station. And this was sort of an ultimate in remote operations. And that, you know, we have astronauts orbiting space into space station, and most of the command and control and monitoring is happening on the ground level. And frankly, the safety priority is the same thing. And we have, you know, as humans, as people, we're our number one priority in their safety. And so they're not just NASA. It's oil and gas, autonomous ships, air traffic control. I mean, there's all kind of work going on in this area that we should be leveraging as an organization. So thank you for your time, and I look forward to the panel discussion. Thank you, Rick. Our next speaker is Dr. Claire Blackett. She is a senior human factors consultant at Risk Pilot in Sweden. Claire has almost 20 years experience working as a human factors specialist in both research and industry, including 11 years as a principal scientist with the Halden Reactor Project in Norway. Claire has a background in root cause analysis for accident investigation with a PhD in this topic. She has worked primarily in the nuclear industry throughout her career, providing human factors and human reliability analysis support to nuclear safety cases, as well as conducting human factors engineering assessments and providing input to event investigations at commercial nuclear sites. Dr. Blackett also has experience in the petroleum, maritime, rail, healthcare, and process industries, and continues to work in all of these domains today. Please welcome Claire Blackett. Thanks, Dave. And thank you, everybody, for joining us today. Just to mention Risk Pilots, since I'm sure that most of you have probably heard of the Halden Reactor Project, so you know about that, but since December, I have moved into a new role with Risk Pilot based in Sweden. Risk Pilot, they've been going since 2002, originally set up as a risk management consultancy, but since then has expanded into including systems engineering and human factors, and so I'm part of the human factors team there. There's nine of us. We have a very strong presence within the nuclear industry in Sweden at the three nuclear power plants, so we have most of us work on site at those plants. And nuclear in Sweden, you may or may not have heard, is going through a bit of a revolution. It's fantastic. In January this year, Sweden announced new plans for expanding the nuclear force. Now, I moved to Sweden in December, and they decided this in January. I'm not sure if it's because I moved there, but you can join the dots yourself. So in Sweden currently, there are three nuclear power plants, Forsmark and Oskarsham, and they produce about a third of the electricity in Sweden, so they're extremely important to Sweden. But in January this year, it was decided that they're going to build more, and so the plans are for building two large-scale reactors by 2035, so the goal is to add an additional 2,500 megawatts by 2035, and the equivalent of 10 new reactors, or 10,000 megawatts by 2045. So it's a really ambitious goal, but they're very, very committed to actually doing this. And essentially they're assuming an electricity demand of about 300 terawatts by 2045, so they sort of acknowledge that between their renewables and the nuclear power, this is how they're going to achieve it. My previous work in the Halden project towards the end of my career there, I started to focus a lot on small modular reactors, and I was very fortunate to be able to work almost exclusively on this topic via the Halden reactor project, and that included being involved in the experimental programme at the new Halden SMR simulator that was installed in 2021. This reactor has the capability of simulating a plant of up to 12 units. We had it set up as a six-unit plant, and this enabled us to actually see how operators can run and monitor and control this plant, a six-unit plant. So we had run three experiments by the time I finished my time at IFA using operators from the US and from Sweden. These are licensed operators currently practising, and none of them had operated in an SMR before, of course, because we don't have any commercial SMRs yet. But we put them in this situation, and we ran them through different scenarios to see how does it look. And the kind of topics that we were interested in was looking at multi-unit operation and disturbances, staffing, roles and responsibilities, how do operators monitor across a six-unit plant, and how do they manage their workload. We also talked about, although we didn't do any experimental studies on remote operation, and very much looking at the changing role of the operator. So very relevant for this session today. In terms of what I saw from the research that I have done before, and then also just thinking about where human factors is going, when I considered what is the greatest challenge or opportunity, again, I landed on the topic of automation as being one of the main challenges. I think that when we talk about automated SMRs, we kind of don't really realize what we're talking about. Automation is not a single discrete thing. It's not just another system component that you can plug in and then say, okay, now it's automated, what do we do? Automation is much more complex. It's much more likely that we are going to have layers of automation within our systems. Some parts of the plant may be more highly automated than other parts of the plant. So we really have to think about the complexities of automation where and how and why have systems been automated and then need to understand and how is the human going to interact with that. We also have to think about is this going to create new types of human errors? Errors that we don't see in our existing plants. And then how are we going to be able to assess those from a human factors perspective, from a human reliability perspective? And then touching a little bit on what Karen said as well, is also thinking about the methodological side of it then. How do we as human factors professionals approach these potential new errors that are going to occur? I agree with what Karen said about we need to be kind of quicker in how we do our human factors assessments. When we look at the world of SMRs, for example, we see that there are so many vendors out there. This world is evolving so quickly. We as human factors professionals need to be able to keep up with that pace. So I think that the methodology, the approach that we use for human factors is also going to be a challenge for us going forward. As always when we talk about human factors, we never have enough data. So how do we generate enough data to be able to validate designs and substantiate claims that we are making on operators? And then also what will be expected by the regulator is another big question that at least I've gotten a couple of answers on that from this week. So in my opinion, those are the challenges that we will face. Thank you very much. Okay, our next panelist, Dr. Paola Savioia is the head of the operational safety section in the Department of Reactor Regulation at the Radiation and Nuclear Safety Authority of Finland. I have to say Paula challenged me with this introduction because it's got the combination of the longest institute names I think I've seen. So if I pause in the wrong spot, please forgive me here with this. From 2001 to 2016, she was a research scientist and team leader for the Human Factors and Complex Systems Team at the VTT Technical Research Center of Finland. While at VTT, Dr. Savioia obtained a master's degree in engineering from the Department of Automation and Systems Technology of Helsinki University of Technology in 2003. In 2014, she completed her doctoral studies with distinction at Alto University School of Science. In her dissertation, she developed a systemic usability concept and evaluation methods to help the development, excuse me, deployment of user-centered design approaches within complex socio-technical systems. She has expertise from several industrial domains, including nuclear power, maritime, metals, and machines, process control, and fire and rescue services. In her current work at STUK, she is taking part in the nuclear energy legislation renewal from the perspective of operational safety. The legislation renewal is part of the Finland's preparatory works for the licensing of advanced reactors. Please welcome Dr. Savioia. Thank you for the introduction. You did great with all the long names. Good afternoon, everybody. It's really nice to be here. I come from STUK, the regulator of radiation and nuclear safety in Finland. Just to give you a little bit of context for what I'm going to talk about, I will very briefly say something about the nuclear landscape in Finland. We have five operating commercial reactors. As of now, four are a little bit older ones, dating from the late 70s and early 80s. But all those older plants, they have gone through some sort of automation, renewal, and digitalization. So the control rooms in even our older plants are so-called hybrid ones containing digital systems and analog systems. But most important experience for myself is our new unit, OL3, which has been in commercial operation about a year now and is actually going through its first yearly outage now as we speak. And OL3 is quite highly automated plant. It's an EPR, 1600 megawatts, and it's located in the western Finland. Then also we have had... One thing I want to point out from this nuclear landscape that since I've been working in STUK, we have had nuclear facilities in all lifecycle phases, because we have the operating plants, the new unit. We also had a construction license review ongoing for a while, for a totally new plant. And then we also have a final repository project that is in construction and commissioning right now. The new construction license application was withdrawn, but from that we still have experience from how human factors have where integrated in the basic design of a new plant. And on top of all this, like David said, our country is going through nuclear energy legislation renewal and this one goal for this whole renewal is to enable licensing of new reactor designs. So let's go to the human factors challenges. I want to illustrate with these two control room pictures what I think is the challenge for us in this period of time. On the left-hand side, there is this old control room from where we as human factors scientists have a lot of experience with. And we know what are the human performance issues and what to look for in evaluating the suitability of such control room. But as we are moving towards this more modern era, we don't have so much experience. We have some research experience, but not real operating experience from these. And for this reason, I think that our human factors methods, we cannot be sure that our methods that we are using right now are such that they can actually point out from the modern control room what are the real threats to human performance. So it's a challenge for us human factors experts to develop our methods in a way that we understand what are the vulnerabilities of the modern control rooms that we are going to face in the near future. And in addition to this, the human factors engineering methods that I've seen being used in the industry are very much focused on the control room issues. Even I put the control room pictures there to illustrate this. But from my experience with the new builds, I've seen that there are quite interesting human performance challenges also out of the control room. Especially in... There are the field operations, but there are also human activities like maintenance and assembly testing, surveillance activities that take place on the field. And what I have seen from the projects that I was... I have been working with human factors engineers have not paid attention so much to these other human activities that are still quite safety critical. So my other challenge relates to this, that do we even understand the role of human performance in the context of advanced reactors? Do we know what are the safety critical human activities where human factors should be concentrating on? Surely there will be those control room issues, but there must be something else too. So I think there are some blind spots that we don't even know about yet. My last slide is a little bit of experiences that we have gained on these other human... safety critical human activities. Some of these were surprises to us and some were something that could be expected. So for example, relating to operating crews, we saw some competence issues, but they can also be interpreted as design issues relating to procedures and administrative procedures. And these are activities that the safety significant is not necessarily direct, but it's indirect, but there still is the safety significant, so human factors engineers should pay attention to designing the tools for different kind of functions at the plant. The last point there concerning emergency management is just an example that other functions, not only the control room operations, might become such that human performance is threatened. Thank you. Thank you, Paula. All right, my turn went up. So what do I believe will be the greatest human factors challenge and opportunity that will be coming to the nuclear power industry? I will start by noting that I did not include in the future in my title here. As I will assert, the greatest challenge and opportunity are here and now. You will also note, let's see, yeah, it's there, that I included a subtitle, a risk-informed perspective. I'm perhaps regretting that I put that in the title now that I see Dr. Apostolakis sitting in the front row. So, but I did plan to qualify that and that it's qualitative assessment. Admittedly, it was hard to discern between the futures that I saw what was going to be the biggest consequence. So I really focused on likelihood. What was the probability of what's to come forward? I need to provide this disclaimer because unlike most presentations where I'm really just handing out the facts of what the NRC has been doing, these are really my views. And so in no way should these be interpreted as necessarily representative of the U.S. NRC. So beginning with the first question, the greatest human factors challenge, I started to think about the trends and topics that are getting the most attention in the human factors literature and in the nuclear domain. And the result of that reflection is what I've depicted on this next slide. Starting at the top, we have remote operation of nuclear facilities. And it was just a few weeks ago, the NRC hosted a workshop on remote operation of nuclear facilities. And in that workshop, we heard a range of plans for remote operations. And in general, I think it's possible we could be seeing proposals for constellations of small reactors being monitored and controlled from a centralized facility that may be far removed from the reactor site. Progressing clockwise, another development, we could see is autonomous operation of nuclear facilities. Facilities that have the capability to maintain steady state operations, handle plant upset conditions, and shut down to a safe and stable state with little to no human intervention. A third development would be the use of artificial intelligence in the operation of nuclear facilities. AI applications are already in development for optimizing nuclear plant design, maintenance, corrective actions, and operator training. We should anticipate that AI will be developed to serve as a decision aid to operators, and perhaps someday serve the same function as shift technical advisors providing on-shift engineering expertise. The fourth development depicted at 9 o'clock on this diagram is human robot teaming. Robots have been used extensively in nuclear cleanup operations, and now free-roaming robots are gaining employment in the performance of nuclear facility surveillance tasks. I expect that successful performance in these applications may lead to future deployment and security in operations applications. Now I could choose one of these futures as the greatest human factors challenge, but the probability or consequences of developments in that area would be open to debate. Rather, I ask you to consider the intersection of these developments. What they share in common might be our greatest challenge. We have... So there it is. Could that be our challenge? The human factors professionals might be surprising for me to say that. Is it a conspiracy? No, it's simply a matter of numbers. The human factors community is small, relative to other engineering disciplines, and extremely small when you consider only those with deep knowledge in the nuclear power domain. At the same time, we are seeing new vendors and innovative designs entering the market. The individuals you see before you as panelists today are what you might call unicorns. They are quite a rare breed. And they are the type of people you will need to move successfully into any of the futures depicted on my prior slide. It is easy to fall into the trap of thinking that as we move forward into a future where operation of plants is less dependent on human operators, that there will be commensurate decrease in the need to consider the role that humans play in ensuring plant safety and human factors engineering. I am cautioning you that such thinking would be part of a recipe for bad outcomes. If you consider each of those futures deeply, you will recognize that they are not really eliminating the role of humans in ensuring safety, but rather they are transforming the when, where, and how. So how do we come to this situation? In short, the old post-TMI talent with nuclear industry experience moved on to other industries that were more rapidly evolving. They have now largely retired, and sadly many I regret to say have passed on. The new talent coming out of universities are being drawn to those industries that are consistently funding research and innovative technologies, such as aerospace, telecommunications, defense, medicine. The flow of human factors talent to the nuclear industry is currently a trickle. And this limited supply could result in advanced reactor design solutions that are delayed or worse ill-conceived, if this challenge is not given due and timely consideration. With that cautionary note, let's move on to a brighter topic. Okay, so what is our greatest opportunity? Let's begin, again, with the same four areas. I believe it was just yesterday that Chair Hanson noted that optimists see challenges as opportunities, so our challenges and opportunities here are the same. And again, we could make an argument as to which of these holds the greatest promise as an opportunity. But as before, I'm going to hedge my bets and look at the intersection of those futures and consider their commonality. Whichever direction or directions one might head, the greatest opportunity is the same. And the good news is that it is here, it is now, you don't have to wait. In fact, if you wait, you'll miss the opportunity. Perhaps I should just pause then for the next slide to build the dramatic tension, but I'm running over in time. So, the opportunity is addressing human system integration early in the design process. I think you've heard that from my other colleagues here in their presentations earlier. So, for those of you who may not be familiar with the term human system integration, I would like you to think about it as one of the primary objectives of human factors engineering. But simply, from my perspective, it's the process of optimizing a design by respecting the needs and capabilities of the human, the hardware, and the software elements of a system such that they work seamlessly together to achieve the system's mission objectives. I noted earlier in this presentation there was a big uptick in human factors following the accident at TMI. Those human factors efforts were important and worthwhile, but they were tantamount to building lifeboats after the ship had set sail. It's not what I would call addressing human system integration early throughout the design process will provide the best opportunity to identify and implement optimal design solutions and minimize rework. As such, it will pay dividends not only during the design process, but throughout the operational life of the plant. It can be your most cost-effective path forward towards safe and efficient plant operations. So this concludes my remarks. And I think at this point what I would like to do is perhaps go back and I know we put up a polling question before. Can we put up the results of our first polling question? Okay, so we had very few that have no idea what human factors means, and it's probably a skewed audience given that he came here to this session. But I'm going to ask if any of my panelists who cared just for those that may not be familiar with human factors, how would you explain human factors to your friends or family as a discipline? Anyone want to take that one on? Well, I'll say, you know, as I explain it to my wife and my family, I usually use my phone as an example. I think we've all of us have been in a case where we have no idea that we want to do something on our phone and whether it's an app that we can't navigate or we can't find out what information or where the information we need is, you know, the design of that interface is where human factors comes in to make things, technology, your phone more easily usable. And so to me, human factors is, you know, designing systems and workplaces and phones with humans in mind, with our capabilities and our limitations and so that's my simplistic explanation of human factors. Good. Thanks. Anyone else want to add or amplify on anything? Rick said? It was a good summary. Good summary. And, you know, I think I've explained it to people in the past in ways that were probably more focused on how we treated here in the nuclear industry because I always made reference to complex technological environments. Human factors really can be applied to, you know, the most basic of tools. We tend to focus on a rather complex tool, a nuclear power plant and so I usually find myself explaining it in terms of the challenges, you know, addressing the challenges of dealing with human limitations and capabilities in a complex technological environment, but it probably shouldn't be thought of as just limited to that. And also thinking about it, we tend to also focus on just the cognitive aspects because most of the work in a control room where we have focused historically has been the nature of cognitive work, but human factors also encompasses that physical aspect of human capabilities and limitations as well. Dave, to that I would like to add that also social aspects of human activities must be considered. So humans as team working and communities. Yeah, absolutely. And we find more and more you know, in our community we try to incorporate you know, at least for example within the NRC our work in the area of safety culture it all ties in as well as other organizational aspects. Going to ask that we put up the next polling question and that was the audience has now heard from all of our panelists in terms of their views on what they considered to be some of the most significant challenges and I heard some threads comment throughout those presentations and some interesting differences but it'll be interesting to see what our audience thinks. So give a moment here to see what has resonated with our audience. And I'll just note that while that's coming in clearly this poll was created before I developed my presentation so none of the options that I happened to suggest are on there so you know I guess maybe that's saving me from embarrassment of finding out that my views just did not at home with our audience but let's see what comes up here where it looks like use of AI is maintaining a lead. So let me turn to my panelists. I guess I'm Paul I'm going to start down here and work this way. What are you thinking of the poll results here? Any reactions to either what's coming in low or high? Any surprises you're seeing? It's really it's difficult to predict what will happen in the future and if any of those those challenges or topics become reality they will transform the way nuclear power plants are operated so they will create challenges for human performance but maybe it is the use of AI is I kind of tend to agree with that it has the most potential for complete transformation of the human role. It's not working? Oh sorry maybe I don't repeat my whole answer. Claire what's your thoughts? Yeah I agree with Paula that it's not exactly surprising to see things like AI and automation coming out on top. I think it reinforces the message that you gave Dave about the need for early human factors involvement. That's one thing that I think we've seen with the rapid increase in the development of AI over the last couple of years is that it's kind of advancing at a pace that is almost outrunning us and it's clear you can see clearly where people have taken a human-centered approach versus where they haven't. There are a lot of issues around AI when it's not designed with a human-centered approach both in terms of how it can be used, how it can potentially be misused and that might not be an intentional misuse, it might just be an artifact of the way it is developed. So I think that if we do start to see AI coming into our nuclear power plants and I do believe that we will we need to think very carefully about why we are introducing it and where we are introducing it and making sure that we understand that operators will actually interact with this and use it as a tool to avoid any potential pitfalls. That's a very negative opinion but it's human factors. But to add a little bit I think it also means a significantly different competence requirements for our operators in order to understand how the AI is working and why the justifications for the decisions that AI might be suggesting. So significant competence requirement changes for operators in any of those but particularly with AI. Yeah, I think the pooling is right on with the AI. One of the principles that we have for our safety systems in our plant is that they're deterministic meaning that we understand if we give it a certain input that we're going to get a certain output from the system and with AI it's not like that. Sometimes the activity or the automation that we get from AI could be unpredictable. My experience admittedly with AI is limited to chat GPT and other interfaces like that but I know that if I ask a question into chat GPT sometimes you get different answers and it's not a predictable response and so as someone who's not an expert in that field that's one question is a lay person with AI would ask and would be questioning is the predictability of the response of AI. Maybe I shouldn't be but I was a little surprised to see on the list or at least the pooling data I'm looking at right now is almost the lowest I think that might be one of the one items that we actually have a general design criteria for our plants is to have a control room that is actually one of our regulatory requirements that appears low on the list so again maybe I shouldn't be surprised but that was interesting to see that people don't see that as one of the top challenges. Those are my thoughts. Those who who know me won't be surprised to know I have a little bit of a different take on this. I would suggest that the biggest impact is the fact that all of these things are going to come together and we aren't going to have just AI and we aren't going to have to just consider AI that's why we have to think about this in a totally different framework. We're going to have all of these things possibly in the same plant and we aren't prepared for any of it but the impact is going to be that's where the challenge is going to be. We have to consider it as a whole and so it isn't just AI and it also surprises me about the facilities without control rooms and I think maybe people think it's because there's no humans in it but it's like the Tesla driving itself when they have to step in that's the biggest challenge too so yeah I think it's the fact that all of these are going to come together and then multiply exponentially so people I think I would agree with Karen's comment there it was one of the reasons why I opted when showing the potential futures or future technologies to show them an event diagram I had to cut my comments and didn't have an opportunity to note that as you look at these some of our can be considered enabling technologies for some of the other futures so I do think that we will see those in combination you know the AI piece was one of the reasons why at the beginning of my presentation I or at the introduction of this workshop I said think about human in the loop as opposed to man in the loop because when you think about AI and it's you know tremendous capabilities the one challenge that we know still remains is that you know how do you build morality you know into an AI and so and that may be one of the things that some of you thought about in terms of you know the difference between humans and machines so I'm going to pause there and so we can move on to some other issues I guess before I move off this topic hopefully you've all noticed that directly following this session and we did this on purpose is that there is a session that will be on AI and so if you have a particular interest in this topic please attend that session we we wanted to make this a broad discussion of advanced concepts of operations not just AI there are a lot of other interesting challenges to be dealt with and they'll go much deeper into this topic in that subsequent session I understand that we have a question from here in the room or comment they wanted to have made so I want to acknowledge Dr. Apostolakis maybe it's not a good question for this session but we keep talking about the AI as operator aid I don't know what that means what can AI do for the operator maybe it's for the next session but we keep talking about it I don't know what AI does so one of the areas where it could be considered an operator aid is as a decision support so if we think about concepts for future designs we tend to talk about them as being highly automated systems and so they will be producing huge amounts of data and we know that machines are much better at crunching data than humans and so that is one area where AI is being discussed is in terms of decision support it's not clear to me like how that will look yet but at least that's when we talk about it as an operator aid that's one type of aid that comes up a lot is a kind of decision support system and when saying operator aid it doesn't necessarily mean the control room operator it can be any human operator who is benefitting from the results of the AI and I guess I would add to that and you think about it as a decision aid so my understanding of different types of capabilities that are being developed would be to aid in diagnostics and prognostics of system performance so you might expect that an AI system might be more capable of picking up weak signals in a system performance that just might seem like noise to a human but the AI is capable of analyzing that vast amount of data and recognizing that there is a anomaly in providing some sort of early warning of a degradation that might not have been picked up okay so I'm going to kind of just move from AI as a specific means of achieving automation to just the question of automation in general thinking about it more generally what do our panelists think about a push for more automation in reactors designs and what might you see as the implications from a human factors perspective at risk of repeating stuff I said already and the things that the fellow panelists touched on I think we struggle to understand the risks when we move to highly automated systems and they don't lessen they just move and they change shape but they don't necessarily know how to assess those to provide substantiation for licensing I think that'll be a challenge I think there is a mistaken actually I've seen both ends of the scale the mistaken impression that it's a much higher cost to introduce automation I've also seen that it's a much lower cost to introduce automation and I don't think that's true I think the cost just moves to a different you know different point and I just think the challenges are different I don't think it decreases the challenges I do think there's a place for automation and I think there are plans especially the smaller plans with the lower consequences where I think it just makes sense to automate I'll say you know based on my presentation of course I am a proponent of increasing the use of automation and you know at the same time understanding the implications and we've seen you know in the industry where direct benefits from the increased use of automation in our existing operating fleet of reactors where we've been able to I've talked a little bit about self-diagnostics where the system can detect failures and sometimes failures that the operators can't even see and we've talked about AI I know we're going to talk about AI more later but AI could even enhance that so we've seen real tangible benefits for that also from an operation and maintenance cost standpoint where think about if you can now credit the system to perform some of these activities you've now reduced the effort not only reduced the effort of performing the test reducing the effort to maintain these testing and surveillance programs and procedures and also free up the operators and other support staff on things that may have more of a safety implication to your plan and so I would offer that up that we've actually seen real tangible benefits today and now with the increased use of automation and I'll also say you know we've been talking a lot about this increased use of automation and you know I've talked to this question to one of my colleagues and it was noted that there is a potential idea to sort of provide some standardization on you know what types of things that we automate and what types of things are best left to the operator so you know in human factors we talk about function allocation where you know we go through all the functions and we say okay this is best for automation this is best for the operator or this is best for a combination of automation and manual control and so perhaps as we you know think about the functions that are consistent across our designs which may be more and more of a challenge as these design changes but as we think through that maybe there's an opportunity to provide industry guidance within our community to think about you know think about these questions from a holistic standpoint and not just from an individual standpoint so Yeah I think that automation is something that is inevitable that will happen and I'm glad that you actually touched on the function allocation part of it because I know I'm not the first person to think this but you know there is that question of just because we can automate something does that mean that we should automate it so I think that when it comes to automation we need to be very clear about what are excuse me our requirements or our objectives when deciding to implement automation so one thing that is discussed a lot in SMRs for example is the opportunity to reduce the number of staff per reactor unit and I often hear that that is discussed as well we can automate the plant so we can reduce staff and it makes me ask the question are we automating to only reduce staff or can we reduce staff because we have automated and I think that those are kind of slightly different sides of the coin we need to be very careful about where we are automating and why we are doing that because it will actually impact upon other systems and subsystems it will impact on the operators we have this idea that you know especially in SMRs because we'll have increased use of automation increased reliance on passive safety systems that if something goes wrong the operators can be hands off for up to 72 hours that's not true the operators will be doing a lot and so I think when we talk about automation in relation to advanced reactors we tend to kind of oversimplify what we're talking about and thinking that well if it's automated we don't need operators and that's very much not the case so I just think that we need to approach it carefully and examine it thoroughly to make sure that we're not oversimplifying something that could actually fundamentally change how operators work in the plant yes I agree and I would like to add to that that as automation is inevitable it has to be and it will happen but it means that automation must be designed with a proper human system integration approach and human factors engineering so that human collaboration with the automation is taken into account and for this we also as human factors professionals need method developments so that we can be properly integrated and understand the impacts of the automation on human activities so I think we need development of methods and approaches in that the comments kind of had the theme of touching upon the issue of function allocation and thinking about this session I went back to an old paper that I thought had some interesting thoughts with respect to where we are today in terms of development and I mentioned that acronym and paper was written by two leading human factors professionals David Woods and Sydney Decker and it was title in part was maba maba or abracadabra and this approach of simplistically thinking about substituting what a human used to do with automation and thinking we've come up with automation that can fix this human problem solved and they described that as kind of substitution thinking that you were just substituting the machine for the human and not necessarily recognizing the kind of downstream consequences of what you've really done now is fundamentally change the system and the way work is done and that there needs to be you know more thinking in terms of just how to make humans and machines work cooperatively together rather than just thinking about this division of labor notion and I thought that was they had an interesting statement in their paper that really struck home with me and so I'm going to quote them they said design concepts represent hypotheses or beliefs about the relationship between technology and human cognition and I thought that was just a very interesting observation we think of design concepts that's solid you got a way to move forward but at the conceptual stage it really is it's a hypothesis it's a belief about how you think the system is going to work and so they go on to then elaborate to say you know such designers need to subject these beliefs to empirical jeopardy by search for disconfirming and confirming evidence I like that phrase empirical jeopardy I guess I just interpreted it to mean validate, validate, validate so we're getting near the end of our panel time I'm going to see if we can take one more question quickly and then we will need to wrap there was a question from our audience this is an interesting one Rick it was directed to you but I invite others to weigh in do you think that we are safer shifting the opportunities for human error from operators to designers well they're both humans right so I don't know if there's any shift going on there I think I understand the basis of the question though but I think it's getting on to the point that I was talking about we have an opportunity to reduce human error and therefore making the plant safer by automating and I think that's true but the second part of that my thought on that and that I presented was you could go too far in my opinion at least too far the other way if you make your system too complex right you introduced the potential of making maybe a quality error or requirements error that goes on detected within the system until it manifests itself as an error so certainly understand that point and I agree and so I think the theme that we've all been talking about is let's not automate just to automate we need to think through the implications and how the operator is impacted to that automation how training is impacted on as we go to advanced systems but that's a really key piece of this one thing I'll say I know we're short of time but I really wanted to share this in 30 seconds this example I've seen operators using newer systems where you can actually where they've received a failure of the system and you can actually see within the control logic of the system troubleshoot actually dive down into the control system of the logic to see where the system may have failed and what I have observed in my personal experience is that operators are very quick to do that as soon as they get a failure to drill down deeper into the digital system to figure out what failed in the system and it was intuitive to them you know and so I think there's an opportunity there to retrain provide more training and more skills to those who interface with the system to be able to interact with the plant more effectively to find these faults Thank you for that observation That's a good point Anyone else want to add on to that I could add that with the commissioning of our new EPR and qualification of our control room operators that's something that we identified that must be paid attention to is the automation competencies of the operators and we really tried to make sure that so I think it's a really important aspect of this higher levels of automation that even when we have the sort of traditional control rooms the operator abilities might and the requirements on their competencies might change a lot I think that's what I meant when I started with is that we need to understand the risks they don't they just move they change the risks to the competency the risk to the training who it is that's carrying the human error risk when we automate so I think that's in line with what I was saying it's this idea that suddenly you've gotten rid of these risks of error and these risks of humans doing stupid things and it's no we're still in the equation we're just elsewhere I'm happy to end on that note but I think that's certainly a a matter that I have been kind of beating the drum on for some time now is for us to keep in mind that as we move to simplified highly passive designs we don't even need to be talking about automation really we just just thinking about designs that have now limited the need for operator intervention to address plan if you're relying on on passive systems or inherent safety characteristics you are then as we've just been speaking to kind of shifting where you need to consider the focus in terms of the potential for human error is it now going to be with the field operator that ensures the proper lineup of that passive system or the individual doing the surveillance of the integrity of the materials of that system or does it go all the way back as this question implied to the design engineer there are plenty points along the way even if you are eliminating the operator from the principle role of addressing plant upsets we need to conclude so I'd like to bring up the slide for our points of contact if they're not able to get it up I will say I'll remind you again that our session coordinators Nieve Hughes Green and Stephanie Maro and myself David D'Zoniers he used the standard format for our email addresses there you go you can email us if there is a need to follow up on this session for any reason we will also be making ourselves the staff and the panelists available to individuals after this meeting out in the lobby if you have a follow-up question we're happy to take those we just need to vacate this room so the next session can come in so we'll take care of things out there there was a QR code up there I think earlier if you wanted to provide feedback on the session or anything to do with the RIC please provide us feedback we're always looking at ways that we can improve these activities so with that I'm going to thank our audience for your questions and attention and please join me in the last round of applause for our panelists