S3E14 - The future of AI: insights from the experts

Speaker A: Welcome to the season finale of Virtually Live, the podcast. The podcast where we get insights from marketing leaders and event professionals. I'm Aaron, video editor at Kaltura. Over the past months we've explored the cutting edge of AI and today we're wrapping it all up. In this special episode, we've collected some of the most thought provoking ideas about the future of generative AI, from ethical questions to exciting applications in healthcare, education and beyond. Enjoy.

Speaker B: So, as an English major, I'm still very much inspired by books. Um, and our leadership team often communicates what books they're reading, what they find, um, really interesting. It doesn't always feel like homework. But the most recent uh, recommendation that came from our leadership team is a recent book called the Geek Way, uh, by Andrew McAfee. And it really does look at the companies that have seen success from leadership from geeks. But I do like the reposition of the geek as obsessive maverick, um, people who really focus on what could be and iterate to get there. And there's a huge juxtaposition. I like the one that they showed between Quibi and Netflix. Netflix listened to their customers and ingrained that feedback in what they do next. And they describe it. I think the description was 70% data and 30% instinct. Quibi on the other hand by far had entertainment visionary but wasn't listening to the market and as a result didn't succeed. So when it all comes down to it, if you ingrain in your culture the data, listening to your customers and using AI even more to accelerate that, to be able to iterate even faster, I think uh, our possibilities are endless.

Speaker C: Something that I learned relatively recently actually was that women's uh, health is only really been around since the 90s, um, you know, so not a lot is known about women's health. So I think that in addition to being cognizant of different cultures and different races and different um, you know, like gender and everything. Right. I think that we also have to be conscious of our different patient, um, uh, population. And so I think what AI is going to do, and I think we'll see this with our large language model that you mentioned, um, before we are taking all of that education from our physician led, um, continuing professional medical education conferences. And we're feeding that into a large language model that will hopefully, um, start creating scripts for us for our patient education. And I think that that will give us a clearer insight into the groups that maybe where we do have gaps, you know, if we're too focused on um, a Certain gender over another, we're gonna see that through. Um, the script writing I do believe and I think that also Mayo Clinic as a whole hasn't really embraced this. But I think um, the idea of where video creation is gonna go to be able to create personalized video, um, experiences for different families, different races, um, I get really excited thinking about the future of that and what that could look like. And that content is geared towards me and my family, um, and looks different for my friend and her family.

Speaker D: I mean I'm going to sound a bit cliche but um, obviously don't be afraid to fail, um, particularly with new technology. But it's changing much more quickly than we can adapt. And so you've got to acknowledge that if you're not keeping pace, don't go down a road you're committed to just because you've committed to it. Um, and we're finding that like on a month to month basis stuff, uh, that we committed to in January, a co pilot implementation has made it redundant and we need to say it's redundant instead of pretending it's not redundant. Uh, and so I said that to me as it relates to Genai is the key thing. Get the tool, whatever the tool or tools are into as many hands in your organization as you can. Because that's actually going to influence where you go not sitting in an ivory tower making a decision.

Speaker E: So we have a cross company working group we've set up um, in the last year called AI at Adobe. And now we all know AI has been around for a long time, but generative AI is really changing the world. And so we set up this working group one to harness that experimentation. So what we realized is employees were excited, engaged and experimenting both in ways that were incredible and we didn't know about, in ways that were terrifying we found out about. And so this working group really has set some guardrails around what types of experimentation we want, what types of software we want to allow in our environments at this point and what types you have to apply to get into the environment. But it uh, really fuels things like hackathons, which as a non tech person I never thought I'd participate in one. But now we do them for here's a problem statement. How do we think generative AI could help solve that problem statement? And so we're doing them in finance, we're doing them in HR in places that we never thought we could. And each business unit has a sort of a lead representation that sits on this committee that is responsible for knowing what experiments are going on responsible for flushing those up and sharing those learnings. And then as we learn, we start to deploy more software. We're also doing things like AI days internally. Uh, legal just did this where they actually held an AI day hosted for the whole company. Talk about where things are going, where policies are, and really helping people understand what those are. We just did one in marketing, which was all about prompt writing and better prompt writing and how to use our own tools. Customer zero is huge for us. As we deploy AI even further in our tools, we're trying to help people understand how to use them. So education, guardrails and lots and lots of experimentation.

Speaker F: So when you think about AI because it's predicated on patterns, it's predicated on what have I learned along the way. If you start obviously with something that's polluted, what comes out is going to be polluted. The transparency along those data pipelines becomes even more important. So when you think about AI, generative AI and large language models, it really starts with do I know where this information is coming from? How will I be using this information? And then from that, from that usage, what are the outcomes that I expect? And do I have mechanisms in place to make sure that if I make a mistake or there are some guardrails that you have governance around ensuring that you can course correct quickly? And so I think, I mean I love analytics. This is why I'm in this business. And I think there are so many opportunities. But the responsibility is on humans to do the right thing. It's not going to be on the patterns themselves. We define what those things are.

Speaker G: Yeah, I think uh, for workday, I think it's uh, embedded in our values. So we have employee, customer success, innovation and integrity and fund. These are the five values that we have. And ethical AI sits between innovation and integrity. And what we are doing um, at workday is that we believe in ethical AI is something that should unlock human potential. It should not replace, it should unlock. So for example, you have an AI companion on your meeting, which document your summarize your meeting notes or minutes of meeting, you still need a human to look at that. And that's where human in the loop comes in. So we need to unlock human potential using ethical AI or AI. The second is how can we really make change in the society using the AI? Uh, one of the example would be that in our product we really don't look at productivity, we look at, okay, fine, how can we make skills cloud? So for example, how can we really make uh, uh, employees or uh companies to really think about skills, what skills they need today and what skills they need in the future. So building that AI around skills cloud as a kind of a visual where we can make a positive impact on the society. The second is around championing fairness and transparency. Uh, we really want to see that we take a risk based approach on the AI. So when a product manager goes in and really want to launch AI feature, he goes through a risk based approach, risk based assessment around that before launching or before even uh, giving that product out and say this is uh, what we would want to launch. And the second is our commitment to data and privacy. So we really want to make sure that we are committed to our customers about your data privacy guardrails and we maintain that guardrails and not violate that by building this uh, ethical AI or AI models in our products.

Speaker H: I spoke on an event uh, last week in Las Vegas and one of my fellow panelists uh, said that AI won't replace you, a person using AI will replace you. And I think that that's a really interesting concept, um, across the board obviously. Um, but I think around events specifically things, uh, thinking about kind of the front end and the back end, we're seeing a lot of applications around the back end specifically video editing, time saving on creating briefs and all that. I did go into work one Monday and ChatGPT worked and then by Friday it didn't work anymore. So there's obviously a lockdown maybe similar to some of uh, you on the panel, uh, where companies don't want you to use certain technologies. Um, but I think that there are a lot of efficiencies on more that backend kind of event, uh, management event, kind of tech production side that will be really helpful. Um, and then also um, kind of thinking about uh, on the front end like streamlining content post event, how do we get content to people from that event faster? Um, and I think that will play a big role too.

Speaker I: I think I'm really excited about the use cases of AI for personalization. I mean again AI is there to serve humans. I think it's funny when we talk about is it going to replace us? And then it's just a bunch of AIs talking to each other. Um, it really has to serve humans. And I think in this post pandemic era, the TikTok era, like attention spans are short and people want to make sure that they're making the most of their time, they're driving impact that they are able to get what they need out of an experience. So I'm really excited about AI. We've been working a lot on how we use AI for summarizing content so people can kind of quickly scan how it can sort of be a little bit of a concierge to help you navigate what can be an overwhelming experience for um, some of our friends. We have up to a thousand sessions that are happening and so we need to make sure we make that very large experience feel small, connected and relevant to them. So I'm really excited for AI to help unlock more of that at scale and at speed, which we couldn't do previously. So that's where I think it's heading. I think that's the, you know, the future of events with AI is personalization.

Speaker J: I think there's ah, education. Everything about education is going to change. If you didn't see the um, the demo last week of Sal Khan from Khan Academy talking with his kid in using that education needs to be completely rethought. And then the other example I see regulated industries. So I know some of you are from banking, from healthcare, but here's the catch. I think they go very quickly in everything that is customer facing. What I always tell them is you can already gain so much agility in your organization and the agility is not only that you save time, it's that you save time from people to do more meaningful work just by doing things that are internal. You don't need necessarily to think immediately on something that goes to the market because regulation needs to keep up. But start small, start internally with the things that make a difference for your team. So I get this question, for instance, all the time, uh, coders. Are coders going to disappear? Probably they're going to disappear, but the coders that now have the experience to understand how you go from something on paper to build something, they're going to still exist. I think there's new jobs that are going to be created. Of course, which jobs they are, I have no idea. I do have one specific pet peeve. This idea of the prompt engineering. You do not need to do a prompt engineering course nowadays. Tools are evolving for as soon as you say what you want, again the critical reasoning, there's going to be a layer anthropic launched one last week in which you just say what you want, they create the prompt and then you get the output. Yeah.

Speaker K: More accurately your decision makings are going to be shorter and exactly like. It resonated with me, the whole decision making thing, because that's where AI for example can make a huge difference. Pre screening Connecticut MRI results and flag it to the doctor. We were talking about a digital twin. So it's exactly tying in to what you said. Um, I mean you talked about cardiovascular and the secondary uh, occurrence. But if you imagine that you have your own digital twin which has all your family history or your medical data there and you present with a new symptom, but then um, we can really have a personalized care and you don't have to go through every single process that you normally would with that new symptom, which might be an incredible strong headache, but, but we already know a lot of your prehistory and we can even stimulate certain um, solutions for your medical care. And then this digital twin will respond to that. That is just such an enormous, um, saving that you can have uh, in time, in effort, in investment, in local presence of hospital staff, that that's definitely the future.

Speaker L: But from a data creation, um, the sentiment monitor, I think as we talked about earlier, uh, with knowing somebody is actually engaged, it's ridiculously hard to measure, right? Unless you're actually putting on their camera to see what they're doing. It's what are they doing on the screen? Are they active in the chat? Is that a good thing? Because if they're chatting the entire way through your meeting, is that a good, Are they engaged? Because I can't listen and type at the same time. So using AI to really prompt nicely, it's almost that Netflix. Are you still there? Perspective, right? But also customizing the content as you go. I think that real time aspect is really okay. What I'm receiving is making sense. How do I then refine that on the fly so that the content coming in at me is what I want, right? Especially if I'm at like um, if I'm at a conference or I'm watching or as a Google Next and we're watching what's happening, right? It's so broad because you're trying to cover it, healthcare, fintech, etc. Right. Like at a red hat level, I didn't have to worry about compliance, right? So every time, yes, I have to worry about gdpr, yes, I have to worry about customer data. But our product's free and open, right? There's no ip, right. We sell, sold free software, right. But now I'm in fintech, I actually have to pay attention to all that stuff, right? So I have to pay attention to credit card information, I have to pay attention to finance. How do I get that snippet? Almost real time. And I think that's, that's The AI perspective going forward is it's less people listening to the sound of their own voice, presenting. It's. It's almost. Here's a collection of data and I'm going to organize it for you so that if I need the right health care information or the right technologies in my hospitals or wherever it's going, you know who I am, you know who you're selling to. And it's the right creation of content for that moment that will keep me engaged.

Speaker M: I think. I mean, again, referring back to the pandemic, it all came from there, right? And I think things have really accelerated. I would, uh, estimate probably five or eight years, certainly on our side. I think for us as an organization, it's not a capability, uh, that we would invest in internally. So what we do is currently we invest in strategic partnerships, depending on what country it is. But it certainly. I talked a little bit about our vision to transform care. It really sits under that. We want to try and enhance health system resilience and capacity. And that is definitely one way to do it. I know that in some of the countries at the moment, we have packages where that becomes part of the offer. So I referred to the, the UM initiative where we were using AI to send out, uh, digital nudges that packaged together, in addition to having access to, um, a virtual healthcare professional as well, is certainly something that we've looked at.

Speaker N: Digital engagement. That's the core of the future of the digital transformation. And second, we heard it also from Marco. We all need to reskill ourselves. The digital twin. It's a new experience. AI is going to be there if we like it or not, if we are denying or if we are angry about that. It's going to be there. The IMF spoke about a billion jobs that are going to be changed in the next six years. And those are our jobs. But it's our responsibility to go back to our organizations and evangelize the change. Change on how we engage and change on how we skill.

Speaker A: Thank you to all of our guests and thank you for tuning in. We'll see you soon. Don't forget to like and subscribe. Bye.