 So while you guys line up on either side here with questions I Asked the panelists to ask questions of themselves, but before I start I noticed a couple of overarching questions that reach beyond Just one of you and one way we have two Students who basically went into business and started up business and I noticed both of you were saying That you were trying to minimize The IP at the university And that led me to think well How did you first of all, you know we want to encourage people like you and your generation to start up new businesses? So is the university at did it help you? Move into the firm or was it something you had to get around like putting all your stuff at the public domain So you didn't have to negotiate Unmute there we go. So Miguel is actually incredibly helpful in supporting our endeavor. I mean, I mean they They basically gave us everything we needed and they never The concern for IP was really our our cautious like our own caution on our part. There was never any mention of it I mean, I'm sure it's important and you know most ventures do fail. So there's there's no point in you know put it put it up a Ruckus, you know, if if nothing's gonna come of it. So who knows if down the lines a Unicorn company comes out of a university what's gonna happen really But I think the approach that Miguel has taken that we've seen so far has been really supportive and helpful And so we're really grateful for that So it's a really great point that you make actually in my case, we didn't have the choice because like The lab that I'm coming from everything is public usually in the code as well so So it was not like a specific question to say like are we going to release the code to to evade the DIP Although you need to protect it at some point or another if it's for investors or others They will ask you like do you possess that IP or do you have a way to protect it? And down the line so in our case we decided anyway, like things move very fast So the stuff that we we were doing in academia Is now like we've evolved from that we've recorded everything I think we have better tools now And that will change also in time So so there was no real point to set deals with the university at that point And now we own the IP so Yeah, there was no big discussion, but it's always the a little bit Stressful or you don't know like if that company becomes a very big company Are they going to sue you or something down the line because like they see the dollar signs So that's still a Big question mark for me. It's not clear exactly well from the path that we've took I think it's we've avoided that but You never know So, let me ask a second question. This is more general is all of you have talked about Some of the stuff being shared and open creating data structures and sustainability of that information and at the same time needing to Have a firm that takes off and has some exclusivity some advantage in the marketplace Where do you learn to figure out that line or are you still learning or where does that information come from? So we're definitely still learning but from from the get-go. We were a very academic oriented group The scientists basically And so we are very much proponents of open science and so our dream really was to make it such that The IP that we create is Not really relevant to our ability to disseminate the knowledge and the science and and the data if possible So we just like if you could just turn like super technical details Proprietary but release all of the knowledge that's created out of your algorithms and the models themselves and the code then I don't know if it's gonna work, but that's I'm really hoping it might Yeah So basically how do you because I mean those little details are what makes the difference at the end like in some in some cases like sharing a model is It can be useful, but it's not necessarily the It's not gonna. Yeah, exactly. It's it's important like at the end. You need to protect it. I think at some point So so yeah, that's that's something that we've on our side Learned in the past few months Doing like exactly trying to figure out what will be the business value What's how to protect our IP how to make value from what we've learned and also from the data as well like How do we gather? public and private data Decilo some of that data that can be very useful and that should be leveraged in academia and in the private sector I think you have a very interesting approach for that You know, I think first of all it's a very slippery slope And we were talking earlier about this if you don't have a framework where you can You can secure the data into a compliant format So one or more people can share it work together then you don't have a framework So it's a slippery slope you're moving forward and you're always looking over your shoulder to see what did I do or what? I did not do so. I mean for example, we've recently at MGH in Boston. We've recently We helped them develop a molecular lab and we set up a small company We're taking that company to market and their objective is to use the IP generator to master general hospital Then allow other people perhaps open science to mix other molecular medicine components Into the mix and then harden the Solution using a company called archer to take those assays and take them to market So sometimes it's an acceleration, but you know, we all started in small businesses I had two to three of them and I've get the scars to bear From starting them because when you're running and gunning and you're just trying to make a living and move forward You know, you don't necessarily have the benefit of this great Legal group that's here. So I think to begin with an auditable system where you can contain mix the data Read-only scratch capability and then preserve it You know preservation of data always helps those who are trying to do the right thing to begin with So I think the kind of system we're talking about is foundational to this So another sort of critical tension that really matters here is where your data is coming from and who it's coming from If you have end users who are consumers You know, you have to be very careful or you might end up being 23 and me selling people's genetic data to big pharma if you are collecting data from consumers, you have to be very very forthright with what you're going to do with it and Especially if you want to share it within an open science framework, you have to be careful about that You know, this is particularly true of neuro imaging data where the de-identification Processes is a sort of a critically important part if you're doing open science if you're sharing or aggregating anything We've had requests from people who say, you know, we'd love to see some of the EEG data that you've got Can you share it with us? And it's an academic lab And so, you know, we're open to collaborating with them and they say I want to build the you know I want to try and train a classifier on Unique brain prints so that I can identify individual users on EEG. It's like, of course, we're not going to share that with you You crazy. Why would we do that? Violate your users privacy so you have to be very forthright with your users and there are frameworks for this sage bio networks is particularly good at this on You know patient-centered or user-centered privacy Making things very easy for the end user to understand understand what they're sharing and then be very careful about what you're doing the sharing what you're doing with the data and You know, I think it there's there's no easy answer to this it's going to be solved on a case-by-case basis for now and The it's really really important that the community engages with that especially, you know People who have a lot of experience in standardizing and sharing data engage with companies that have this increasingly common problem or challenge Yeah, I'd say the other thing that So that goes along with that You know with that standardization and pooling of data is that As there's more and more to work with as it becomes more and more of an issue Privacy-perferring software can become much more important. I think it's it's going to be a growing field That's going to really accelerate the development of all the other fields because it's going to allow access to a lot of data and Those who can show that their systems really do protect that privacy We'll have an edge in terms of access to data In terms of open sourcing I think there's always two things to look at and the way that you open source data and the way that you open source tools and other Software are two very different things. I think it takes two different policies and approaches Towards them, you know, as you said that the trust that you have when you're a data collector is very important Once you lose it you it's very very hard to regain it And what we what we've noticed is that being a trusted data partner on the market today is very very valuable So that's one area where you want to be much more careful than let's say your open source tools in terms of open sourcing tools And you're you're a software. I mean there's always a trade-off between where your expertise lies and where Other innovations are happening in other labs and you want to be a part of it and in order to be a part of it one You have to you know be a Good community member, but the other thing is if you take one version of it And you go off on your side and you sort of fork it and start building on your own You're not going to be able to take advantage of a lot of the innovations that are happening anyway So I think very much case-by-case. I think it's very much a difficult thing But it's about recognizing where your particular value lies and how to protect that while taking advantage of the the other innovations that are going on and I think AI being a Field where it's very resource constrained at the moment knowing how to apply those tools Even if they are open source or having access to particular data, even if the models are open source That can be your competitive edge that can allow you to to contribute the rest back to the community Actually, that's a great point So I was going to sort of come off of that now she talked about a point that grain rays as well Is that one of the things that we tend to downplay in doing this is how much effort we put into actually developing the platform So our own sort of expertise and personal skills that provided that platform And you can obviously reverse engineer a lot of software packages But actually figuring out how to use them for specific applications longer They can't succeed where they don't succeed it's really worth a lot And that's really where you get the competitive advantage and the companies We have talked to as the virtual brain founders have been very excited about that particular opportunity Because the soft the solutions itself are out there ready They can get from a paper no problem there But it's actually knowing where it works where it doesn't work Learning how to use the software in the conjunction with AI platforms for example That's stuff that we've been working on we can take that knowledge and actually help build the platforms together And that's really a huge advantage for businesses So thank you for that one of the Concerns when I talked about it, but we generally know the cost of developing a drug is increasing and people are looking to neuro Informatics and other tools to help address that but though a lot of the costs are in the clinical trials or something else So are what you're producing? I'll be the cynic here. Just added costs to the system or they actually You know do you see an opportunity here for this technology to start driving down the cost of actually doing these trials and bring A new drug to market second, thank you You know I think There are so many components to clinical trials and to research and to clinical Precision medicine by the way, I think behavioral health is the ultimate use of precision medicine There are so many pieces where one has to go out and or the drug companies spend a fortune to go out and to develop a cohort And the rest of it at the end of the day the reason we're doing all of this ourselves is to accelerate Into clinical practice the things that we're all learning so we've worked for example To be able to kind of back-end feed some very very large Healthcare systems with clinical research data and since there is no ICD-9 or E&M code we kind of wrote web interfaces drove it into the back end and Allowed to have reporting within the EMI or another system So it was unified clinical research management system and healthcare system if you do that all of a sudden you're in control of the data And you can reduce cost you can take the control out of these people who are trying to develop the cohorts and adding Extraordinary cost to clinical trials so you can't use data until you can manage data The objective is to manage it and then there'll be some private Apps and you know people who develop these applications Deserve the recognition and monetary enhancement that they get It's not a battle. I think it's a combination of IP open source and open science and That's the mix that we really have to reach and then let's accelerate these things in the clinical practice We all know people practice medicine the way they were trained to practice medicine The only way you could escalate it is evidence-based data that drives them to do the right thing so I'm gonna talk about more Alzheimer's application or Alzheimer's clinical trial but the basically the there is ways to Reduce the cost for the clinical trial or fast-track a little bit some of the the process Basically the main cost in clinical trials is in recruitment screening people to make sure that they They have the right properties for the trial and more often than other especially in Alzheimer They recruit people that don't have the desired quality Like cognitive decline. They should cognitively decline in the time of the trial If they don't or like if they stay stable They will just add noise to the the clinical trial resulting in increased cost So if you're able to target the right population that will kind of decline or reject the ones that will that will remain stable That's a great way to ensure that if the drug is effective It's gonna go through the whole process if it's not then it's not due to the Wrong selection of your subjects or too much noise or a Simple size that is too small for the heterogeneity of your population So yeah, there is ways to to reduce the cost of clinical trial I think quite drastically and if you do that earlier on in the screening process Then you also like save a lot in the the most costly Modalities that you will do at the end of the process So, you know another another way you can think about cost reduction is by increasing the Resolution of your measurements or at the quality of your measurements by do potentially doing more regular measurements with at-home technology wearable technology That costs a lot less than a laboratory visit So if you can collect data remotely for a clinical trial, you can you can potentially At least achieve some of the outcome measures some of the health measures that you're you're looking for Remotely in a much lower cost and much more regularly so you can you can actually look for where these changes are happening What the time course of them is? Longitudinally A good example of this is say sleep measurement if sleep is one of your outcome measures or if you're running a sleep clinic I don't know if anybody here has ever been to a sleep clinic But it's an awful experience and sleeping sleeping in a polysomnography system is terrible Not only are you you know sleeping with wires all over yourself and accelerometers and respirometers? But you basically can't move at night so people don't sleep well so you're actually getting a measure that's not very indicative of any kind of pathology associated with sleep and you know There are a lot of reasons to believe Randy was telling me last week that that sleep is a is an interesting indicator potentially of early cognitive decline Biomarker for Alzheimer's, but it's also you know hope for a host of other reasons It's a really interesting thing if you can make something sparse and wearable that someone can wear at home and sleep multiple nights in then you can get sleep data EEG EMG heart data breathing data a whole bunch of other things all on a Longitudinal basis that you can average over multiple nights or you can look at multiple nights in the time course of this and see You get actually more robust data at a lower cost without having to without without having the cost of a laboratory without having the cost of administering all this equipment and having and potentially you know marrying that with Robust AI classifiers for or sleep classifiers or eat data classifiers you can do a lot of the data scoring automatically and and remove a lot of the human element from from from data scoring I Want to get down to a more sort of fundamental level on that as well. So one of the things that in my perspective at least that's Hasn't really been considered is whether the theoretical framework underlying clinical trials is really taking into consideration how complicated if you will and complex the system is we're trying to cure and and I think We've been sort of using those the fundamental rule that is going to be a single cure for a single disease and that's most times wrong and Sort of looking for that magic bullet looking for that one thing that's gonna change everybody's life is not a good way to proceed with clinical trials So the alternative framework is to actually go back and think about how does the entire system work? And it's actually a complex system. It's a very nonlinear system We had sessions today talking about dynamical systems and talking about the complexity and emergence and so on and so forth And that skill set is actually quite useful to think about how things that happen in your brain Relate to things that happen in your body of things that happen to happen in your environment to your genes and so on and so forth So you think about how to use that knowledge then to to to guide your clinical trial to do a bit more work on personalizing and putting precision medicine That may change in fact how we do clinical trials because we're using much more information about the individual Instead of selecting only a small number of individuals ultimately for clinical trial You can look at everybody and actually make use of these wonderful frameworks are putting together for both AI for Biological modeling systems like virtual brain that taking a consideration the complexity of the system try and use that to really understand How these different systems to interact provide better solutions that ultimately will help the individual as opposed to helping one person out of a thousand So one part of the puzzle is we want more drugs and we want them cheaper and we want them sooner But the other part is on the innovation side, so at least most of you other than Randy are in firms Most of you are in Canadian firms How do we keep you here? So we see that the normal trajectory for a Canadian company as you develop your product with lots of Public money to support you at the early phases you get some private money you prove proof of concept and then You sell to the US or Europe or something. What can the government do? to keep you here I would say the first thing and you know for element one of the things that's been one of the drivers of our growth in Canada has been the talent pipeline, so Continuing that that training continuing to foster an open environment between The universities and the firms here. I think is a very important part of it I think you know, we're never gonna have the Venture capital or at least for the future. We won't have the same venture capital Abilities here as we do in the US And that's not a problem that throwing money at will necessarily solve so it's great to get venture capital from From the government, but it doesn't come with the same networks and expertise that a Season venture capital firm will bring to the table along with that money I think in terms of AI one of the things that will really help Canadian firm succeed is if we find ways to structure data trusts or You know data banks that are available to Canadian Emerging companies, so if we find ways to leverage the data that the Canadian government or you know Canadian people operating on Canadian soil are creating whether it's through actually taking government data or creating You know sort of a data tax on companies operating here and making that available to companies in a way That's obviously privacy preserving That'll help the Canadian operations of these different companies take off And I think that's one of the ways that we really have to be forward thinking It's not a way of operation that's really been done before But I think in order to really compete on the global stage. That's one of the things that Canadian companies are going to need You know, I think that's a great question and Canada generally Quebec specifically has attracted Tremendous industry here and many of our partners, you know Google Microsoft to name a few Are here and they're here for a reason because they have access to to brilliant people. They have access to data So part of the initiative that we're working on is to try to keep things that are positive in Canada it is part of our alliance We're bringing Intel into the mix. We're bringing Microsoft into the mix Which is which is working on very interesting AI issues not so much the actual AI capability But the massive parallelism to be able to pull that data in so at least it's our objective to try to work Within Canada and keep things in Canada. We're looking at companies and working with companies here one that was That was displaying outside is you're on digital pathology which worked with Alan and Catrin and your like Germany when we did the big brain data set. So we're trying to take that and project it Expanded first within Canada. I mean one of the things will do one of the good things about Dell is we're a go-to-market company We we don't compete with our partners. We try to have a joint go-to-market to drive them So one of the things we would do is come up with a simple story Drive a simple story to our sales reps have them tell the simple story to their customers and try to drive and propagate Systems that are developed in Canada within Canada, and then obviously Elsewhere hopefully to return revenue to Canada So yeah, I mean one of the things I think we have to do is just start believing that this can be an industry This can be a big industry. We know what the economic costs of brain disease are going to be if they're not already The solutions to this are going to be incredibly costly and incredibly profitable for whoever comes up with viable businesses to serve these needs So, you know, we've got Dell here telling us how awesome neuroscience in Canada is two days ago at the International Federation on aging it in Toronto There was a fellow from the Dementia Discovery Fund who said yeah I'm taking meetings all over Canada because you know, this is such a special place in neuroscience You really punch so far above your weight If we don't we should almost be embarrassed if we don't become the global cluster of Neurotechnology and of brain health companies We've got all of the advantages, you know academically We're there scientifically that we're there when you look at the work that's being done at the neuro and at OBI with the brain code database in aggregating all of this patient data from all over an entire province with a single-payer system We really should be able to pull this off And even venture capital even access to capital is getting a little better Even if it might never be as good as it is in the u.s. I think if we can believe it then, you know It'll start to actually happen people have to be willing to take a few risks Students have to be like looking around at different career opportunities creating roles for themselves in the private sector And this will just it's not gonna happen overnight, but it will happen somewhere And there's every reason why it should happen in Montreal in Toronto Yeah, I mean I concur with the other ones The the data is very important and to facilitate access to data and that like duplicate the work Let's say if I go to one hospitals to get some data and then another one or even like in different in different subfield of a specific hospital We have a great cluster academic cluster in Canada and in Quebec and in Montreal and And we all want to transit that that research to to some application that may be at some point game changers but we need to a compare to To take care of those of those new initiatives and provide them with the the right support data is one of them a big one actually and I mean the u.s. Market is much bigger, but we have this If we can support and give the data the right data and the right support to those company I think it's not gonna be a problem to they don't need to Even if they have money or capital from the u.s. It's not a problem They can stay here in Montreal or elsewhere and even if their market is in the u.s They can still stay here as long as there is some incentive or some justification like if the talent pool is here That's that's one of the big incentive and the other one. I think is mainly the data Okay, well, I promise that you get to you would have to ask questions of each other I gave you lots of time So who wants to start? You can be to a particular person or the entire panel Well, they're not lining up at the microphones If we want to go to the microphone go to the microphone and then we'll go to you guys and no one come back here So I entered the the whole just a little bit apparently after you started your talk and I'm puzzled by this Expenses for creating IP is going higher and higher and the profit is going lower and lower I can see that this happens in a drug manufacturing a but I don't see why it would happen in Computation, you know compute a computer software AI, etc. Because at the end It's computers that do not cost much and it's mainly the manpower. I don't see why And I believe you mentioned that this happens in all sectors not only the drug Manufacturer maybe you can explain that Yeah, it's not that profits go down. So individual firms aren't necessarily doing worse. It's that The the cost of doing the same degree of research if you want or the same contribution is going up So as long as you know individual firms will succeed Within that by by spending the money and we'll get it out But overall when you look at across the economy We're not getting the economic growth out of our investments in research And that's what's driving Governments to invest in research which is in the end they're hoping for some type of economic growth That is declining that is whatever we're researching is not contributing in the way that That technology in the early 20th century contributed to economic growth So we're failing to translate it as part of it But the research is becoming more complicated It requires larger staffs and and is being sustained so far because we're putting more money into it The question becomes when that money dries up, you know, and we just had the budget increase a lot for at least public sector research When we go back in three to five years and tell the government again There's done enough money into it at some point they may say no and what's gonna happen to the industry So we're not at a crisis point today But the data is showing that we need to find ways to reduce the cost And we're hoping that open science and informatics and so on can help us drive down those costs and be more productive So it's more of a warning sign Because obviously there are a lot of companies that make money now So it's a counterintuitive story I'm telling you because you look around you hear about the iPhone you hear about all kinds of stuff It looks like innovations everywhere But it's actually contributing less to the economy than we would like and in the health care sector we obviously see this with Increasing percentage of our provincial budgets and federal budgets going to or these provincial budgets going to health care Again a sustainability problem We'll have to see in the longer term that it's happened across most industries Frankly, this is too new a sector for anybody to make any firm conclusions about this sector Right, you'd need a longer history to be able to see what's happening compared to others So it might be an exception in this sector, but it's still an overall problem But the other thing about this sector is it's adopting open methods right from the beginning and so perhaps it will Overcome some of these problems because it often if we look at the Montreal community The reason we have such a vibrant community is because we were open at the beginning that brought people in and Del gave the industry some of the things they wanted. So maybe this industry is doing the right thing But we won't know for ten years until data starts accumulating. I Think a couple of general comments about your question What does all of this technology bring to us? How does it help us and I guess first of all Clinical trials are typically incredibly inefficient engines The pharmaceutical company asks a very very limited set of questions of a data set and if the drug doesn't work Then they essentially put that data on the shelf Increasingly we're in the position where they can share their data Why would they do that? Why would they give it away to the other guys? Well, the other guys are putting data in too So everybody can win if you have a slight shift in your mentality that we can explore pre clinically With a lot pre-competitively sorry with a lot of data that we would have otherwise had to find ourselves So I think that accelerates the process. You've got lots more people asking questions of the data Look at me Instead of three people asking data of that data set in a closed sense It's been made open and hundreds indeed thousands of people have explored that data set with all kinds of findings So that's an obvious case in point the second point I'd like to add is that Traditional models of clinical trials are very much stratified We have to make everybody look the same so we can test the impact of this drug, which is you know It the drug has been applied to left-handed white males. It's not generalizable beyond that as A left-handed white male. I think that's great, but it doesn't really help other people Now you put it out there and people can start to look at not tightly stratified data. They can allow masses of comorbid Patients to be Studied in such a way that you can explore much more subtle aspects of Comorbidities in a way which the typical clinical trial doesn't allow and The the CCNA the Canadian Consortium for Neurodegeneration and Aging Does exactly that I thought that it was a very bad idea when they first decided to throw all the patients into into the mix Now it turns out that that's actually a great Vehicle for AI and precision medicine and more sophisticated analytic techniques to tease out those comorbid those Mechanistic components rather than sticking everybody in a common label. You've got Alzheimer's disease You've got vascular dementia those things increasingly don't apply anymore and we're going to look to AI to give us a much more mature Perspective on the disease process Both of those things type of data and the analytics you apply Start here in this kind of form Thank you. Yeah Great panel. Thanks a lot and so I had a question about the legal implications of AI with respect to liability So as AI technologies are sort of coming to market in a serious way There's major questions related to who's responsible if AI fails and as health companies who Whose technologies impact actual patients? How do you think about liability legal liability with respect to AI? Is this something that you're concerned about or are you just waiting to see what happens with respect to courts and What laws we pass in the future? So it really depends on your claims basically if you want to have like an autonomous AI that will diagnose you without any intervention of a medical doctor then Definitely, this is something that you need to consider very strongly because you will have repercussion because you're the last line basically of defense before the diagnosis Most of AI applications are assistive tools to provide more information to the physician And at the end the decision and the final diagnostic is done by the By the the physician. So in term of liability You circumvent a little bit that part that said it's not an exception. It's not a good reason to do a poor job or have Algorithms that actually Have like very bad results like you still need to validate that Make sure that they do generalize on many type of population As on what's mentioning to have a diverse pool of subjects that are very heterogeneous that Like you make sure that you don't have a false positive and stuff like that But that those things can be measured as long as you have the data to measure them or you have access to that data Yeah, and I would add just more generally, you know outside the scope of AI and health I think there's a lot of There's a lot of conversation about what that means in terms of liability There's a lot of sort of doomsday scenarios that are being thrown out There's a lot of talk about you know killer robots and you know if a robot does this I think we tend to attribute More personality to to AI systems than they actually have in the wider community And that's something that we have to sort of fight against The fact is that most of the time These questions will be solved either contractually or By policy so you know a really good example is no fault insurance in Quebec in Canada, so People had these questions about you know if a car is driving and it hits someone who's responsible It's just much simpler on a societal site to societal level to say Okay, this is how we're gonna deal with it as long as there wasn't negligence on the part of the the AI developers and as long as You know as you said you're doing it responsibly. You're using the proper data. There is licensing You know there's licensing regimes in place to make sure that no one's just launching a bunch of cars It's a societal shift. That's happening We have to deal with it from the policy standpoint and there are places where that you know policy isn't there yet But I think that the large-scale deployment of these Technologies also isn't there yet and we just have to make sure that that goes hand-in-hand But it's not gonna be one of these scenarios where no one knows how to address their liability and no one knows Sort of no one knows who's gonna be ultimately responsible the legal principles we have in place still do apply for AI solutions There's a question over here and then After that, okay, this one's yeah one issue that I haven't heard brought up Everyone's commented over how important it is get to get access to the data But I'm a researcher in one of the institutes in Toronto I spend huge amounts of my time interacting with lawyers on data transfer agreements and Worrying about why the REB at that particular institution desperately wants to add a sentence to the consent form It's got a little better in Ontario with clinical trials Ontario But we have different legal frameworks for privacy in each of the provinces in Canada So we desperately need in my opinion and I'm curious to see what the panel thinks the government to actually get its act together and Fix the legal environment in which we're operating them because we seem to be way ahead of Where the law is and the law is way ahead of the people often we end up interacting with within the institutions In terms of what they know about what's allowed versus what we are actually trying to do And it's not just within candidates between countries to with with Europe Canada's reviewing its privacy law and we've got the two levels I mean one of the things when I mentioned open sign a signet is we're trying to Figure out Can we simplify? The I mean there are two parts one is getting the consent of the patient But then there's the sharing of the data itself So even if we get the right type of consent doesn't mean we can make the data available And what happens if in five years we discovered? Oh, we should have done it a little bit differently We don't want to have to re-consent the the patient we have to find other mechanisms So Bartha Knoppers at the Center for genomics is you know, one of the things they are looking at is that but I don't know if other people want to And you've got a federalism issue, which is not resolvable, but I don't know if I would say, you know, I definitely understand that concern the fact is most Most technology companies today do operate globally. They'll collect data At least somewhat globally they'll have global markets the standard practice is to comply with the highest standard and We're lucky in Canada that for a long time we had one of those higher standards So we got used to being a little more protective of data, which made it so that when GTB GDPR came into effect Most firms here were already Pretty compliant with that higher standard the that one of the issues is that I think as a profession Privacy law is something that was not really taught thought about practiced in a Really meaningful way up until recently and so I think like what you'll see a lot of people sort of being really hesitant to make changes to the form or Insisting on inserting this language that they got from their outside counsel that they paid a fortune for and don't want to go Back to is a lack of understanding of the interplay between These laws and these new laws and the technology itself And once you understand that full data flow and how it works as a lawyer It makes it much easier to be a bit more flexible and as a profession. We're not there yet, and I think that's something that we have to work on Yeah, I mean you can graduate law school without taking IP let alone knowing anything about consent laws, so So you may not be the right group of people to ask this too But one of the biggest problems with ensuring that data are that are generated inside of universities Are available is the sustainability of these data resources, and this is not a problem that governments have solved It's not something the research community has Figured out themselves because you ask any Individual researcher if they want part of their budget to go to infrastructure. They'll say no so Industry on the one hand Can benefit from this data and clearly, you know We have to do things a little bit better in the academy to have that happen But do you see scenarios developing where industry can help with this problem because it's really a global? It's a it's a global issue. I don't know of any field that is not facing us That's just a great question and I think of all the things that we're going to try to do Through this partnership with McGill is exactly that how do you How do you sustain a model so you can retain that data for long periods of time? How can you secure it? How can you audit it? And how can you have people come together to share data and then preserve that data set? that's at the very core of what we're going to try to do here at McGill and my belief is is that if we do this properly in the long run, it'll be Significantly less expensive There'll be higher utilization of the it that you have and there'll be a much better outcome I'm sure and I talked a little bit about this earlier I Think if we can build the it systems to be able to to to adequately secure and sustain the data and then we can have Issues around ethics and data sharing Oval in data gravity overlaying that will have a unique combination. So that really is why I'm here I think that's the number one issue Which will accelerate all of the things Yeah, and the and the idea is to not do that at all The idea is to just have it in a containerized environment, which is owned by the people who own the data But you can Microsegment that little environment so that small environment is owned by who owns it Not by industry. So clearly the objective here is not to To to own a controller data But to allow the people who have the data have interacted with the data with other people to be able to prove They did what they did and then be able to keep it themselves either Individually or institutionally that's really the holy grail and that's where we're going So another answer to that is that if you look at the incentive structure of For for data flow or for you know the creation of tools Ultimately the solution of code sharing was not solved by an open-source nonprofit initiative It was it was solved by github and then they sold for eight billion dollars and you know there the Eventually, this is probably going to happen to the the there's every reason to believe that this can happen to the Standardization and storage and maintenance of of data You know something that venture capitalists like to say I listened to probably too many of them But they one of the things that they like to say is if you're a hardware company You know you you're you're worth two times your revenue if you're a software as a service company You're worth five times your revenue if you're a data flow company and you're doing AI and your data And you've got a lot of data coming in you're worth ten times your revenue So there's every reason to believe that that the standardization of data Will be incentivized to happen in the private sector and that you know that rather than trying to charge you for storing data They'll give it to you free and then do a bunch of machine learning and make you dependent on them And make everybody dependent on them You know we've we've thought about you know, how do we maybe we don't just make our cloud We have a poster on our cloud architecture and how we do data science What if we just open up our cloud to everybody who wants to store EEG data and then we figure out how to make better tools Then we are the ones who have all the EEG data Wouldn't that be cool and it's cool from a scientific perspective, but it creates some interesting perverse incentives You know and I and I and I'm a I'm a proponent of open science There are probably lots of people out there who are thinking about this with brain data and with pharma data Who are maybe not so? Benevolently motivated I guess That's exactly right. I mean that's true of all of the consumer all of the big tech Yeah mentally if you can't control audit Maintain and secure your own data that we're going to be talking about this 10 years from now and people talk about the cloud They talk about all these other issues Once it's gone, it's out of your control and the game is over So fundamental basic tools that allow control security auditing your own control of your own data That's where immediate focus has to take Yeah, if I could just point out that Canadian government is currently reviewing its data strategy So given that you are data consumers, you should pay attention This is being driven by concerns that our data is being handed over to non-Canadian entities Specifically around Toronto But the government is actively looking at this Europe is ahead of us in terms of thinking about this They've developed the open science cloud So whole e infrastructure designed to support data without having The worries about control. So do we want that do we want more of a hybrid? These are all open questions Canada in terms of policy development is considerably behind Most of our competitors And so, you know, one of the open questions is what can government do here? And I think any infrastructure is clearly Part of what the government ought to be doing not only to provide the data But also all the governance mechanisms around it and it hopefully helps all some of the consent issues and I would add one thing that isn't probably isn't thought about very often in these circles, but That this is one of the things that competition law should be addressing There was a review recently in the last few years of the competition act and in the first drafts of it They focused a lot on data and you know players you have Dominant market positions in terms of data and what that meant and whether there should be positive licensing obligations on that type of player How that could be done? So that there was sort of a beginning of conversation about that To no one you know to any no septic surprise. Anyway, the next version of that paper did not contain that I think there's a lot of there's a lot of lobbying around that and being you know, as we just said being aware paying attention having a conversation about the importance of Competition in the data field is something that's very important Okay, we've hit Well, we've hit our time limit. I'm looking to our our organizer. I follow instructions One more question, okay, we'll take one more question because I've been given permission to have it asked Okay, how about I asked both of you ask your question and then we'll put it to the panel, okay Well, I'm kind of surprised. They haven't heard the magic word blockchain yet Not I'm not mocking as I've worked in it for many years. It has to do with things you're talking about data provenance data security data accessibility You can sort of think of a blockchain as this this Global database that nobody owns. It's like a whiteboard in the cloud where people sign documents and they're indelible and just a little example for example Now if I wanted to get my dental records Instead of trying to go through this siloed proprietary dental office database. I could have my dentist Sign and encrypt my documents and put them in the cloud. No one can see him But I mean, oh it can interpret them but me I know they're there if I want to sign them over to a new dentist. I can do it at any time This really I mean this there's a there's a whole world of possibilities related to blockchain That I think people should maybe maybe you are I don't know but I haven't heard the word yet Well, thanks for the layup on that because I agree blockchain is Perhaps the most important technology out there Blockchain and Splunk both So we're working with blockchain There was an indication on one of the slides that I passed We're working on blockchain both in the health care life sciences business as well as the financial business It's probably the most important technology out there and very important in consenting as well You can track all of that across the enterprise so at the end of the day the kinds of things we're talking about Containerizing data and using blockchain to overcome a lot of these problems is likely the answer and some of the big Payers in our insurance companies in the States are really beginning to aggressively use blockchain and Splunk for the very reasons you're suggesting Thank you There was one. Yeah Hi, so thanks a lot to the panelists. It's a great discussion I'm following on really from the question that came from somewhere over there about liability and there was an allergy made with Self-driving cars regulations and this is a very, you know, emerging area and so on but I'm really curious about and maybe I suppose maybe Sonya, but I guess all you guys can speak to it as in terms of informing the clinical decision-making process with new AI based technologies or Non AI based told technologies whatever but something that's not gone through the the kind of canonical scientific validation process What is what it how do you? In terms of the discussions you have with the clinicians how how do you kind of propose to inform and improve the clinical decision-making process and then just a kind of small follow-on to that is if there's a clinician who has like 10 AI companies who all have a product that is claiming to You know give them some improvement in their diagnosis Is that kind of an issue that could be coming on the horizon if they they get kind of saturated? Is this something that is you know, you've given some thought to Yeah, good question. So we actually do enforce the the traditional Clinical validation pathway, which is why we we started doing those clinical trials like I had mentioned earlier and Was the other part of your question? Okay, yes, so a lot of it actually isn't a two in clinical decision-making support systems a lot of it is not actually related to the advancement that AI can bring but rather enforcing measurement based care in clinical practice and especially in psychiatry making sure that you know clinicians are reporting in and and collecting standardized reporting measures of their patients in a way that is aggregated somehow and and then used to inform decisions and a lot of it then goes into the interpretability of ability the abilities of interpretability of on are able to interpret AI decisions and What sources of data and evidence base that an AI? algorithm draws upon to Suggest a certain treatments and to make that available to the clinician in an accessible format You know we all we all talk about big data big data AI and all the rest of it small data is also a very good thing you know population health and other things really bring forward issues that are invisible and Particularly in behavioral health and I call the behavioral health people the invisible patients because there's not necessarily in the record any kind of designation that really suggests that they've got a problem, but there are other markers such as obesity or or other issues as my diabetes and the comorbidities that indicate that that person needs behavioral health Help to get them off the couch to get them back into a productive environment. So, you know We're working with Hopkins using their their ACG's Grouping technology to be able to do exactly that kind of work So there are so many simple things here that we can do Without leaping towards really the exotic where we can really make a dramatic change. I think with that we are Preventing you from drinking So thank you to my co-panelists and I'll turn it over to Elena to Advise you where to get those Yes, so again huge thanks to all of our panelists, I think We still get good stuff left though for the day poster session starts now there's drinks and Canapes or something along those lines outside. So please check out all of the posters and demos And then we have the social tonight information should be in your programs book so after the post poster session, please join us at the venue For more drinks and more food and live jazz. So thanks for today you guys