Transcript: Steve Anderson, NarraViz, on Reliable AI — Interview with Alexy

Uh my name is Steve Anderson and I am the founder of Narrowaviz which is a aentic data visualization platform that allows people to get to insight within seconds. Uh I've worked in a lot of different industries, most industries from the NBA to uh sports to government to retail to financial services. And the most interesting one was with um here in California where we used statistics and machine learning models to influence health care public policies that have saved the state uh a lot of health care expenditures as well as uh make people healthier. So and that's what I think AI is. AI is something that creates value to to people. Yeah. My definition of reliable AI is AI that can do a task in a way that's reliable, in a way that is secure, safe, in a way that creates value for humans. And that's that is the biggest thing that I think uh reliable AI is uh again create value to humans. A human is in the loop and it's transparent. It's explainable. and and so forth. And I think that's what reliable AI is. Yeah. So, so what do we need to build to make AI more reliable? Uh I think that we need to think about it from a couple of different dimensions and I'm thinking about these dimensions on the fly. Uh one dimension again is the people dimension. uh how is it going to affect uh uh people lives and it need to have a positive effect on people life uh if you cannot think about how it's going to impact people maybe it shouldn't be done um even if it's a B2B type of thing that it still has to have some effect on people technology wise uh technology wise is just because you can do it don't really mean you can do it how does it fit in the overall tech technology landscape that's out there. So if you're at a company, how does it fit within your tech stack? How does it fit within your AI strategy? How does it fit into your overall corporate strategy so that you are hitting the outcomes that you want? Uh regardless of of of what you're trying to do. So a lot of different dimensions, people, technology, corporate strategy, uh of course we needed outside uh AI as well in a social environment. How how how do we uh tackle some of the world problems that we we have from better patient outcomes to uh environmental and and so forth. So how do we define those problems and then come up with solutions a reliable AI solutions to tackle those? So that that's that that those are uh the dimensions that I think about. I I think it's going to keep augmenting uh what uh what humans do. I think that it's going to make us more efficient. It's going to create more opportunities. Um, I think it's going to what I what where where the area that I want to see it advance is is continue to advance in uh the healthc care space like like literally improving people health. I got some there there's some healthcare organizations out there that doing some really great things around it. Um UCSF is doing some stuff. Northwell in New York is doing some stuff around early detection of cancer. Um so I my hope is in five years we see more of that because the scale at which you can make an impact is is astronomical. Like for example, Northwell is scanning 10,000 images per week of people abdominal images to see if they can detect uh uh cancer early and they have uh tremendously impacted patient outcomes, positive patient outcomes. So if we can go that route that that that that would be uh re really awesome and just instead of just doing things that that that that we we don't need. Um, as for a lot of people always talk about jobs and so forth, uh, how it's going to impact jobs. At the end of the day, you you're going to need you're going to need people. Um, and, uh, and I just think that, uh, that as technology advanced, um, uh, how do we prepare society for the new skills that's going to be needed? and and that's going to be a corporate issue because they need jobs. That's going to be a governmental issue because um uh that's gonna be education uh uh from from grade school to college have to figure out how do I infuse a uh AI skills into everyone. So I think it's going next five years going to be us figuring this thing out and see seeing how we can improve people out and see how we can bring more people along the journey because today from an AI adoption standpoint not enough people uh have the skill set needed for for what's what what what's to come but overall people people going to be still needed.