Exploring the Future of AI

So as Raj introduced the topic that I'll be speaking on today is exploring the future of AI. And, you know, as you can imagine, putting together an event like this, there's a lot of close collaboration between Vince team marketing, team Analyst, advisors, you know, determining some of the topics that we wanna bring to you guys today. So there were some, of course, some very enthusiastic voices. Yeah. What's the future of AI. And then I think, oh boy, future of AI. Great. But I'm an Analyst. I can figure that out.

So I'm gonna take you on a ride with me for how we can come close to thinking about the future of AI. So First I'd like to establish some common ground here, because of course, whenever we talk about AI, you know, you get three people in a room and all three people are talking about something different. So not to throw out yet another definition of AI. But I think something that we can typically all agree on is that AI is a field of fields. And now this is something that Thomas Edison was purported to say about electricity. Meaning that electricity is in itself, a field of study.

You could focus very closely on this, but it's more of a catalyst which opens up so many more fields of development of innovation and ultimately changing the way we live and the way we do business. And so AI is very much the same. There are of course, many different domains that we simply just stuff under this umbrella and call it AI. But what we need to be focusing on are what are the fields, which are being opened up in front of us. So that's part of the question on what is the future of AI.

And so what I would like to do with you today is to look through this question from a few different lenses. So first of all, wondering what the future of AI could be through its purely technological advancements. And we could spend all day, all month talking about that, but it'll be very brief. Well then take a look at a huge impact on what the future of AI will be. And that's of course, through national and regional strategies here, shaping the investment, going in behind this. And then lastly, we have to consider if AI is going to be part of our daily lives.

And in many ways, it already is that often comes through enterprise and adoption. And so then we consider the different ways that an enterprise may choose to go about this and bringing AI into their future. So first of all, technology advancements, I'd like to show this trend compass, first of all, and now we've got two lines here. We're gonna focus on the blue at first. And now what this is showing is the development of technologies over time.

And here we're measuring the, the very, very broad AI umbrella term, everything stuffed underneath it, you can see of course, nothing new it's had a long history, took off, had a large leap in hype with the advent of cloud computing, but we're moving slowly towards maturity. There's a lot of room for improvement and growth here. And we expect that yes, there will be improvements in maturity here also increases in hype, but we're nowhere near the end of this climb yet.

And of course we can consider under this umbrella term fields like NLP national language process, national language processing, oh my goodness, natural language processing, machine learning, robotics, computer vision, things like that. And so what I've done again, we could talk far too long about this, but I've just pulled out a few things, which I am more interested in looking at in the next year. This is something which has caught my eye. I don't know much about it at the moment, but this is where I want to pivot my research towards and stay abreast.

So I'd also, I'd love to hear your opinions, what you're interested in following in these technological developments. So first AI chips, woo for edge computing, capable of bringing this machine, learning to the local device. And so what I've heard this being used in at the moment is autonomous vehicles, but you can imagine so many more other implications and uses of this. So curious about that, what I've heard quite a lot from vendors as I'm doing briefings with them is natural language query.

And this is a means of, of course, making machine learning more closer to and interactive with an actual human using this filling off of what Mike you were saying earlier, that we need to have ways where this, these calculations are understandable and relatable in order to be used and in order to be trusted. So lots of times these can be put into chat bots into platforms, into BI tools. So this is interesting and I don't think we've reached the end of where this can go.

And finally, of course, intelligent IOT, you can imagine connected devices, realtime monitoring, analyzing, and coordinating their own behavior. So again, these are areas where I'm curious about, I'd like to know more about, and I'd be of course, interested to hear than of you are experts out there on this love to talk, or if you have other areas that you're curious about, let me know. So then moving on to another lens here where we can predict the future of AI. So that's of course through national and regional strategies.

So this is of course an abbreviation and I would, again, would love to hear your thoughts on this as well. So what I've done is I pulled out three different countries and regions. So the us China European union, as these three are very active in the AI space in different ways. And I pulled out a few different factors to compare against and just give the overall feeling of how these different regions are approaching and supporting AI development. So the first is cohesive strategy. And so you can see this is the bar graph on the farthest left and by cohesive strategy.

I mean, looking at a document or a, a strategy, a statement from a federal or a regional level, which is clearly communicating an intention and steps forward to achieve that intention about AI. So you can see us on the farthest left. There are statements at times they are contradictory. Unfortunately from what I have found would be curious if you have found something different that there is little cohesive strategy here that is being communicated to the public in a coordinated way, the EU, this is a step up.

We could say there are many documents coming from a regional level, which state very clear intentions. Now what I find missing here are what comes after the intentions. So implementation and clear steps. A lot of times this is left to individual member states that's fully acceptable, but that does mean a slower rule out. So that is an implication on the future of AI. We can't say this is good, bad, negative. It's simply what's happening at the moment, China.

Yes, there are very clear statements. And in fact, they're broken down into two, five year plans with very clear intentions and achievables and responsible parties. Okay. Sounds good. Moving on is regulation.

And again, this could be argued in different ways. Is regulation an enabler or is this an inhibitor? And you could, there are good arguments, both ways would love to hear and have that conversation with you as well.

Again, the us, there are a handful of executive orders and whatever regulation there is, it's typically geared towards weapons or autonomous cars. So there isn't regulation on AI as a whole. So there's that the EU we can see has the highest score here.

And yes, they are the closest to bringing a regulation clearly on AI out, there's a public draft available for you to read online at the fastest. It could be past next year coming into full effect in 2024 then, but it could take longer then taking a look at China. This depends on your interpretation then while there is very little legally binding regulation, there are very clear plans, which are in some ways being carried out as regulations. So depend on your interpretation. There investment then is broken into federal and private sector or public and private sector.

Then you can see us little public sector investment, but very high private sector investment. And if there is a strategy coming from the us, it is this, it is a market driven approach towards facilitating innovation startups. And in bringing in expertise for AI development, the EU has, is a little more weighted towards public se sector investment rather than private in China as well. And then looking at the, at the last factor here, commercial freedom of market driven investment can see the us is performing quite strongly here.

So again, would love to have a conversation about this. If you're interested, I'd love to hear your opinions, but let's keep moving forward. So then we could consider the future of enterprise adoption of AI. And so when we consider again how AI will make it into the lives of individual of individuals and have an impact, hopefully positively on their lives, on their work, it's going to be coming through enterprises. And so then to consider how our enterprises and why would they, what are their different options for bringing that into their organizations? So we have our trend compass here.

Again, this time we're looking at the red line and this is looking at something called AI service clouds. I'll come back to that in a moment, but we've determined that there are, are generally two ways, very broadly speaking, that enterprises would adopt AI. They could opt for a point solution, and you could think of this as in AI company, they specialize in a type of technology. So being very focused on language solutions or on classification solutions and then even a step further.

So being focused on chatbots for a particular use on image classification for healthcare, and of course even more specific for a particular disorder for a particular machine to be used that. So these are prebuilt solutions they're ready for customization, but really they are designed for the use case, which you as the organization already have in mind clearcut. Now a lot of the potential that AI brings for organizations is that it is a very flexible tool.

Any problem that you can imagine, maybe not any but many problems that you can imagine in your organization could have an, an AI solution there, not to say that that is the best or most appropriate, but there's a lot of flexibility here. And it requires a lot of creativity then to problem solve your way towards a new and different way of running your business, whether it's redesigning the business model, alternating processes so that you operate in a different way. And that's where the concept of AI service clouds is coming into place. These vendors are typically also offering cloud services.

And so they bring the cloud computing that you need for training and deployment. That's also something very important, but they help you design and they give you the tools to design your own ML pipeline. So machine learning, operations Mo ops. So it goes from beginning to end connecting your own data sets to auto selecting a model, kind of the ML inception here using ML to build your ML for your solution all the way through the end. Typically with governance modules included.

So this is the red line here on this trend compass being compared against the development of AI technologies in general. So AI service clouds obviously come on the map much later, again, slowly growing here we are in 2021, we have a moderate market size. There are good number of vendors out there, but it's something fairly new, not as well known and sometimes not always sought out by organizations, interested in getting involved in AI, we project. This is going to take a larger leap and maturity and grab, grab the interest of people as well.

So then some recommendations here for organizations that are interested in taking a step towards bringing AI into their organizations. However, that may be first of all, is really prioritizing and making sure that if AI ML is the, the right solution for the problem you're trying to solve, that must be absolutely clear first. And that's the assumption that we have to move forward with for the rest of these. You should also make sure that this is an alignment with your digital transformation goals.

If it's AI simply for the sake of AI, it's going to be a superficial change and it's not going to bring the, the transformative change of going digital that you may hope. So really make sure that this is an alignment with a larger digital strategy. Some technical prerequisites, there has to be a very strong data governance groundwork here.

So, and this means constant care of the maintenance of data sets. And this is going to be actually a, a ground level capability for all organizations working towards a digital transformation. Cause this is going to be the, the quality of your business intelligence, the quality of your AI, everything there, the quality of your data, privacy programs, it all starts with a good data governance considering your staffing. It would be great to have full staff of qualified ML engineers out there, but that may be very difficult.

So what you're likely going to need to do is have a couple thought leaders here who are willing to bring the rest of your team along for the ride. This is another really interesting contribution of AI service clouds is that they work very hard at democratizing AI and making it possible for team members from different departments, with different levels of training to meaningfully collaborate on the same project, which is a huge part of AI ethics as well of being able to get other perspectives towards solving a problem to hopefully not marginalize some users or have some unintended effects.

So there's that. And finally embarking, really making sure that you align this with your strengths, including making sure that you have adequate data for your own model and training that, or if you need to be working with more global data sets, that is something you have to determine. So then surprise question. We've been looking at what is the future of AI? We also need to ask what does AI mean for our own future? And so to look at that, no surprise. If we start at the top left corner, we move around clockwise. There's no surprise here.

If we bring more AI into our lives, there will be a higher dependence on data. I hope this is not a shock for anybody, but what this will mean is that we will need more targeted security around our data and our data sets and making sure also that the resiliency of our data flows are there as well. This becomes a form of a supply chain here in order to keep everything moving as expected. So this will require very targeted security. And this will also require ethical persistence.

Now it's really excited or it's really easy to get excited about the implications of an AI solution for this problem that you've been just laboring over for ages, trying to get this figured out and working better. It's easy to get focused on the positives and lose where some of the negative impacts may come from. So be intentional about including a multi-stakeholder team, as well as including end users and building that communication and that transparency and that moves into the last thing. And if we're talking very metaphorically here, then AI is a form of communication.

It is transforming data into decisions and insights. So taking something less intelligible to somebody paying attention to something, a clear message. And so if we take this even further, if AI is communication, we do need to be building better communication within our organizations, because this is something which will likely cross the different silos. It will also be of very high importance for your management board to be informed about, especially if there were an incident or a question they have to be informed very quickly and be able to SPL speak clearly on that.

And this also travels down to the other end as well. You need to be in communication with your end users and also determine what level of information they should have about individual decisions and how to invite them into that process.

Like, and as well as keeping it secure. So with that, I very much thank you for your attention. I welcome your questions, your conversation. Thank you.