Remarks by President Velshi at Innovation Showcase
May 9, 2019
– Check against delivery –
My name is Rumina Velshi, and I am the President and CEO of the Canadian Nuclear Safety Commission, or the CNSC.
It is my great pleasure to be here with you this afternoon to participate in this intellectually stimulating event – the first of many more to come, I hope. I am hearing great things about the hackathon currently underway and think the judges are going to have some tough decisions tonight deciding amongst all the great ideas and innovations. I hope that the healthy competition embodied in this event will encourage ever bolder and more innovative ideas.
My intent over the next 15 minutes or so is to share with you a little bit about my organization and then consider some of the innovative technologies presently impacting various industries, how we deal with regulating innovation in Canada’s nuclear industry, the potential impacts of not doing innovation right, and the importance of collaboration.
The CNSC and its Priorities
I am sure almost everyone here is well acquainted with the CNSC, so I will just note that we are Canada’s independent nuclear regulator and our mandate is safety. Over 900 staff across the country diligently ensure the licensing decisions and conditions made by our seven-member Commission are implemented and respected at all times by licensees.
The CNSC’s four key priorities are 1) to have a modern approach to nuclear regulation, 2) to be a trusted regulator, 3) to maintain our global nuclear influence, and 4) to improve our management effectiveness. In addition to these priorities, I have a personal commitment to promote careers in science, technology, engineering and mathematics – or STEM – for women and girls. As an engineer with almost 40 years in the nuclear industry, I have seen steps in the right direction to break down barriers for women in STEM.
But there is still a long way to go. I challenge everyone in this room to be an ally and take action to make positive changes toward gender balance in our workplaces.
Our priorities will ensure that we continue to use science-based, risk-informed and technically sound regulatory practices in all that we do. They will also help us strive to be appropriately agile and respond to an evolving workforce and industry. Being agile will position us to be able to evaluate the regulatory implications of new and innovative nuclear technologies while taking into account uncertainties.
As you know, innovative and disruptive technologies are driving changes in many industries at a pace never seen before, and we can expect the same for the nuclear industry.
Innovative Technologies Impacting Industries
In some industries, change is moving quicker than regulators can, forcing some to respond in haphazard ways. Think of Uber, Lyft and Airbnb, and the disruption brought to the cab and hotel industries.
Some local governments have thrown their hands up and stood aside while these companies have moved into their jurisdictions, while others have fully embraced them, and yet others have used whatever tools they have to try to control them. I believe we need to find our balance in this exciting era. It is an era where change can happen so unexpectedly and so profoundly that it leaves many of us overwhelmed, whether in our personal or professional lives. We now do our taxes, drive our cars, and get medical diagnoses with the aid of assistive technology and machine learning programs.
I can ask out loud any question I can think of and have my home or phone assistant provide me an instantaneous answer. Those answers may not always be useful or correct, but the continuous improvements in response time and content are amazing. How long, I wonder, before we no longer have to ask a question out loud, but merely think it? In many industries change is coming quicker than we are ready to admit or consider. Innovative technologies and approaches are already being used in the nuclear industry in Canada and around the world, including 3-D printing of parts, inspections using drones, and the use of predictive analytics for maintenance of components.
Regulating Innovation in Canada’s Nuclear Industry
As the regulator, we are taking the steps necessary to be ready:
- We continue to modernize our regulatory framework to make it as technology neutral as possible.
- Our regulation is risk informed, so we use the best approach for each circumstance.
- We have a vendor design review service we offer to identify any fundamental barriers to licensing in Canada. That service is presently being used by 11 small modular reactor vendors.
- We have recruited over 70 new graduates in recent years to account for our projected attrition and to ensure we transfer valuable corporate knowledge. We are making sure we have the right skills and knowledge on hand or in the pipeline.
But, given the innovative changes we are seeing, for which there is no history or expertise for us to draw on in some cases, are we in nuclear, both the regulator and industry, truly ready? As the regulator, we have no desire to be a barrier to innovation, but we will never issue a licence for something that we cannot satisfy ourselves will be safe. To me, that reinforces the importance of keeping us informed as early as possible and doing absolutely all the work necessary to be able to demonstrate that the proposed applications of innovative technologies or approaches will not compromise safety.
Innovation can never be seen to undermine safety principles, particularly those that have been developed from decades of experience with both research and power reactors, such as separating process and safety systems. And it means, above all, never trying to minimize or mislead on any issue identified that could potentially have safety consequences, no matter how seemingly inconsequential. Finding that balance is important if we want to be sure to avoid potentially disastrous applications of novel technologies or innovative enhancements to existing technologies. If not done right, innovation can have disastrous consequences in any industry, even those whose technologies have been around for over a century.
Potential Implications of Not Doing Innovation Right
Think for a moment about what recently happened with Boeing’s 737 MAX airliner, and the potential analogy to small modular reactors. Two planes fell out of the sky and 346 people died. Investigations are ongoing, and my comments are based only on what I have read in the media; some have suggested that there may have been a link between an innovative, automated system added to a decades-old technology and the interface with the pilots. There have been many stories that Boeing, the licensee, might have played too significant a role in the certification process, principally because the Federal Aviation Administration, the regulator, did not have the capacity to appropriately assess the technology. The entire fleet is now grounded pending improvements by Boeing that will satisfy regulators and airlines. The United States Congress, FBI and Office of Inspector General are all investigating the certification process and inspector training.
I have asked my staff to monitor this file to see what are the relevant lessons learned for the nuclear industry and our regulation of it. So, what issues should we consider in light of these tragic accidents?
When we are dealing with new technologies, or innovations to existing technologies, are our current processes and approaches appropriate? Or do we need to rethink how we do things and find new ways to approach problems and challenges?
When we consider some SMR designs, they too are bringing innovation to decades-old technologies. These will be first-of-a-kind projects for Canada, and the public will expect and demand that they be demonstrated to be safe. And never forget that the onus remains on applicants to demonstrate to the Commission that they will be able to safely operate any technologies they propose to use.
I also fully expect that someday, perhaps even whilst I am President, a licensee will request the use of an automated artificial intelligence-based system in a nuclear facility. And licensees themselves may not fully understand all potential ways that this type of system could fail, and will have to make hard judgements about the amount of testing and simulation necessary to validate it. Some disciplines are already wrestling with the "black box" problem where the response or solution from a neural network to a given problem or question is correct, or seems correct, but the logic, steps or processes used for the answer are not known or understood.
For example, the medical industry hopes to make use of big data for personalized diagnoses and treatments. If a neural network uses that data to find complex, implicit relationships to make diagnoses and recommend treatments that doctors do not fully understand, will doctors feel comfortable implementing them?
Will patients? We will likely see this same debate play out in the near term with respect to autonomous vehicles. Who we hold accountable and how for errors or accidents from these automated systems are likely to have a major impact on the timing and extent of their implementation. Will we be able to accept the implementation of an automated system in a nuclear facility that makes decisions based on input and reasoning that we do not or cannot fully understand? I am not trying to discourage innovation; as I said earlier, we encourage and support it. And innovation done well can result in important efficiencies and cost savings for the industry, and more importantly, enhance safety.
There are, of course, administrative implications to consider as well. Presently, under our cost recovery model, we are able to recover costs for regulatory activities only from the time we receive an application or a request for a vendor design review. This model does not support the regulator proactively developing capacity in anticipation of licensing requests for innovative technologies. We need to look at alternate funding models to be ready to regulate innovation. The big question for us, then, is how to prevent the regulator from being a roadblock to innovation. Regulators are always looking for safety improvements and view with suspicion any changes without a clear safety benefit. To be more accepting of novel solutions, regulators need to embrace a risk-informed approach, as we have done, and not only ask "What is the risk in this new solution?" but also, "How does it compare with the risks of the current solution?"
But... because these technologies are so new, so different, what more should we do?
I think it is clear that collaboration is vital, whether domestic or international, and among vendors, operators, regulators and governments. Safety is a shared goal in all nuclear countries and that provides valuable opportunities to work together to ensure safety, particularly when dealing with innovative technologies. We work closely with international peers through organizations such as the International Atomic Energy Agency, the Nuclear Energy Agency, and through bilateral and multilateral relationships. Through those relationships, we are often able to leverage particular expertise that we might lack.
However, there may be times when you bring something to us that is so innovative or unexpected that the capacity to respond to it is just not available. We need to acknowledge that, and you need to consider that. We do not wish to stifle innovation in Canada’s nuclear industry, so it is important that industry work with us and keep us informed of their work and intentions as early as possible and every step along the way. I have tried to think of a current example of a regulator that is ahead of the innovation curve and is known for its nimbleness to respond to innovation, but must confess that I have not been able to. It may have been possible 20 or 30 years ago, but given the often limited resources regulators have and the pace of technological development, regulators are at a distinct disadvantage.
To sum up, innovation is impacting industries around the world and the nuclear industry is no exception. Oftentimes, these innovations bring attention to questions and concerns that none of us yet have adequate responses to, or have even thought of. But we can never let innovation give rise to overconfidence or inappropriate risk taking. Think back to the Titanic and the safety innovations it incorporated, which so emboldened the designers that they did not include sufficient lifeboats and led the captain to venture into territory best avoided.
I would argue that Canada’s nuclear industry is as safe as any industry in Canada, and we hold you to the highest of safety standards, even though those often do not seem sufficient for some Canadians.
These are exciting and promising times for the nuclear industry, and I encourage you to continue to be bold and innovative. Your challenge is to be able to demonstrate to us the safety case and your ability to appropriately manage any proposed application of an innovative technology or approach.
We support nuclear innovation, and are taking steps to ensure that we are ready to regulate future innovative technologies or approaches you propose. We must ensure at all times that we proceed safely and smartly into our bold new future.
- Date modified: