Keynote Opening, GenAI State of the Union 2024 with Mark Hinkle at AIE 2024

Welcome everyone to the Artificially Intelligent Enterprise online. I am the conference director Mark Henkel, and I'm very excited to kick off the inaugural, um, a IE online conference. This has been a, a very, very exciting year for ai, and I appreciate all of you coming to hear what we have to say.

Today. What I'm gonna do is I'm gonna give you a state of the union, and this is sort of where we think generative AI is today and where it's going. We have lots of exciting speakers, um, from all over the industry, everything from executives to AI developers to desktop productivity experts in generative ai.

So, uh, should be an exciting day and we're happy to have you. As I said, my name's Mark Henkel. I run an AI consultancy called Parity Labs, and I've spent the last 25 years in emerging tech, and I help people understand and use artificial intelligence.

And as you all know, since you were, um, subscribed to the newsletter when you joined, I, I publish a weekly, uh, newsletter to that effect. Before I go into the State of the Union, I'd like to thank our sponsors. I'd like to thank sny, Okta Techstrong Research and Sisense for helping to keep this a free and accessible event for all of you, and I really appreciate their support.

So let's dive into some things. Let's start with the numbers. Artificial intelligence is gonna change our world.

7 trillion, um, by 2030. That is a lot of money. That's a lot of improved productivity.

That's a lot of new services. That's a lot more change that we're gonna see. And what's what's important to understand is that generative AI is going to augment our lives.

Um, there's a lot of of fear around that, and that fear is because we believe that it's probably going to displace 85 million jobs. And notice I said displace, and I didn't say eliminate because that will also result in the creation of 97 million new jobs. And we see every time a new technology comes around, whether it's the electric motor, the PC desktop, the automobile, that there's a period of disruption and then followed by a period of productivity.

And that's what I'm most excited about, and I hope you are as well. We also know that we think that it's gonna improve our businesses. And as in a poll of, uh, nine 97% of business owners believe Chat, GBT will help their business and improve the productivity and creativity and quality of the work that their, um, their productivity workers are doing every day.

So why is there such a, you know, uproar around generative ai? The uproar is really around, um, the fact that for many years there have been machine learning and, and artificial intelligent software that we have used, but it's never been as accessible as it is now. And I just mentioned chat, EBT and chat.

EBT is probably the best demonstration of all time of how this artificial intelligence can be used by virtually everyone on the planet. The enhanced accessibility is, is the number one thing that has sparked this. But the other thing is that it's actually provided massive efficiency and and productivity gains.

So we're seeing people that are seeing 2, 3, 4, and 10 times more productivity because they're able to take the natural language processing capabilities of CHA GBT and apply them to their everyday work. And the third thing is just the technology in the last six or seven years has moved at a rockets pace. And the, the, the thing that is interesting is we've gone from the ability to analyze data and draw insights with it, and that's sort of the machine learning.

But now AI is actually taking those learnings and generating new types of content based on what it's learned. And that's, that's the real benefit that we're seeing as a society. So the kinds of things, and this is, this is the stuff that you will see every day, um, generating original content.

We see a lot more content online that is being generated on, on ai, but we're also seeing companies that are doing this in mass that are generating documentation, uh, creating reporting that are doing other drawing insights. And that generation of original content is, um, the transformative effect of it. We're also at the beginning of automating complex tasks.

And, um, you'll probably hear about artificially intelligent agents. Um, one of the most, uh, common ones we talk about these days is Microsoft copilot. But these are in their infancy.

They, uh, they're able to, to automate simple tasks, but as these models get more complex, they'll be able to automate even more and more complex tasks so that we can let AI do things that are time consuming and less valuable. And we can exercise our creativity and our insight and our judgment on top of what they're, they're, um, creating. On top of that, they're creating massive economic value.

Um, we will see companies that are going to see an exponential group increase in growth. We'll see new companies and new technologies that are gonna enable our existing companies and organizations to, uh, thrive. I talked a little bit earlier about how many jobs would be affected, um, as far as displacement and creative, but it will probably affect 80% of the US workforce.

And at least 10% of our tasks will, uh, be affected by LLMs in the near term. And that doesn't need to be a negative thing. It means that we'll have more time to work on the things that are, are more important for us as well as, uh, the 19% of those folks that are, are, are impacted, uh, significantly.

Um, they'll see 50% of their tasks either augmented or affected or replaced by ai. And if they're like me, I don't have enough time in my day to do all the things I want. And those are the things that I look forward to.

Since adopting generative AI in my, uh, consulting practice, I've seen at least the three times productivity from before. I still do a lot of the same things, but a lot of the, um, low value, time consuming tasks have shifted to me being able to think deeper, uh, be more creative and spend more time on tasks that are more creative to my bottom line. So if we look at the skills, and this is from the World Economic Report, um, that, that people are looking for in the future for smaller companies, they're, they're still looking at at things like big data and other things, but for companies with over half a million, uh, employees, AI is the top priority.

These, these skills are the skills that we need to, to move in the future. And, uh, it, it's, uh, uh, 15th was where the big AI and big data ranked, uh, for 2023. And for 2024, it's, it's moved up to number three.

So we, we, we can see the shift as we survey these end users of technology and where they want to go as far as a time. So we talked about the tasks, but now we're gonna talk about time. And, you know, time is our most valuable asset.

That time is where we spend, um, creating new, new products where we're talking to customers, doing all these sort of things. And, and of the working hours, 40% of our working hours across industries are gonna be impacted by generative AI and large language models and large language models. You know, we'll, we'll have many talks today that'll, that'll dive deeper into that.

But the, the number one thing that you want to remember is these large language live models are like a librarian that has read every book in the library. And when they encounter new data, they use that information that they learned from the library to imply inference to new situations they encounter. And that's, that's what we're, we're really doing, is we're, we're adding a assistant or the term of our today is copilot that can sit there and provide inference for data that we encounter, um, based on this large purpose of data that it's ingested into, its its memory.

So think about the reason it's gonna impact us is because there's 40% of working hours are actually going to be 40% of our hours where we have a dedicated assistant to help us, or dedicated assistance. Uh, and we talked about how it's gonna affect jobs, how it's gonna affect talent, despite the fact that it's, it's, uh, affecting how we work. Most, um, of our business leaders still think that there's a talent shortage and the number one place they see that is in cybersecurity, followed by engineering and ironically, creative design.

And that's one of the, the areas where there's been a lot of debate is, will AI try and replace our artists and our musicians and our writers? And I, I think the act, the actual answer is that these models are additive to their capabilities. I still think that human creativity, the insights we draw, the judgment we exercise is gonna give the human side of things and advantage.

We're just gonna become more productive, more efficient. One of the great uses for generative AI is to just brainstorm and ideate on new concepts. But at the end of the day, it requires a human being to understand what those concepts have value, and it just sparks our creativity.

It doesn't, um, overshadow it. So let's get into some more numbers. And these are the numbers on how AI is really improving certain areas.

Um, if you look at, uh, development software development, the, uh, um, GitHub copilot for Microsoft has been one of the big success stories. The, they estimate that it's included, it's improved their, um, ability to deliver software, um, by 55%. And really what what they're talking about is GitHub copilot is almost a auto complete kind of solution where it can help you write the documentation for, for software.

It can help you with sub routines. It can help you, you know, write, write low value code. But the, the developers are still going to put that together in a way that makes sense.

And that's, that's sort of the, the, the developers looking and debugging and doing things that they're very good at while benefiting from this copilot that can help them write their code faster. Just think about auto complete in your email or in your word processing programs. Those are the kinds of things where if you can do that the whole way through your code, it's gonna be much more effective.

Uh, another place is video editing. 90% of the time that that video editors spend on cutting and pasting and trimming and clipping, you know, you're going to have a smart agent. And the one that I, that this 90% came from runway is runway ml makes avidity generative AI video tools.

Um, they were used for, um, the Academy Award-winning, uh, movie, everything everywhere, all at once, a couple years back, um, for the rock scenes, not so much as far as the artistic design, but just finding the right places to clip and combined editing. Another great use of this is answering customer support requests. We've seen as much as 45% reduction in human answered support requests.

'cause a lot of those support requests are just queries for information. So if you look at the, the, uh, support agents that are on, um, websites today, you're gonna see them become more and more adept at discerning what the question is and giving them better answers. And, and that's really allows you to, um, use your human power to, um, do the soft skills deescalate customer support requests, make them much more effective.

And finally, uh, we saw 79% higher quality in AI chat, chat rated physician responses. And I don't think that this is, and this is all information from, uh, CO two, which is a private equity and venture investor and surveys they ran. But what they're seeing is that these chats are able to provide more time with the, the, the, uh, patient and help to discern what the all the symptoms are.

So, um, you still need the physician oversight, but part of the, um, discovery of ailments and symptoms, um, can be offloaded to a smart AI doctor's assistant or doctor's copilot. So if we talk a little bit more about the, the skills around this and where we're going is, if we look at the World Economic report, um, the forum report from last year, the skills that we're looking for today are not necessarily AI and big data skills, but really the creative thinking and the analytical thinking that, that we need to make sure that we're providing human oversight for these systems. So while AI skills are very important, there's no replacement for good judgment.

And that's what we can pull from this. Sta this is one of my favorite stats because I think this makes it real. So if we look at, at the fact that, um, technology has improved productivity over time, and this is a coalition of, uh, data over, you know, from the US Bureau of Labor and Statistics, that that really shows certain times in history when we've had a productivity boom after things like electricity and personal computing.

So if you see the two green lines on the right hand side, you can see, you know, where we, we had this huge technology, electricity to our homes, and it was very disruptive for a very short amount of time. And then there was a big peak afterwards. You know, the same thing happens when the desktop computer comes along and, uh, becomes in a fixture in every home in late nineties, early or late eighties, early nineties.

And you see that once again, we see a big spike. And that's what's probably gonna happen when we adopt more and more generative ai, is that we're gonna be a little disruptive, or maybe a lot, and I don't wanna downplay that, but as a whole, it brings a huge impact to the productivity of our society. Without these, uh, improvements in productivity, we would've had greater poverty, uh, less food security than we do today.

And these problems haven't been solved, but honestly, they're quite a bit better than they had were centuries ago. The other thing to think about is, and I noticed I had mentioned that there's a disruption and a, a, you know, some jobs will go away, but a lot will be, um, improved. So you can see that, that we believe that occupations that existed in 1940 didn't, or, um, now, you know, areas of growth today.

So you can see professionals, managers, clerical, and admin. We can, we don't know exactly what those job titles and job responsibilities will be, but I can tell you that if we look at history, they will be less physically, um, demanding. They will be, provide a better standard of living, and they will give us a better and higher level of information availability.

And, and then that comes down to the humans that decide that they're gonna act on that. And I think that's a great opportunity for us, despite the fact that we're gonna see these improvements, there are some legal and ethical concerns. So the, probably the number one concern around AI is personal privacy.

Um, and this was the same concern that we had on the internet. This is the reason that we have laws like the GDPR and to make sure that our data stays private as well as, you know, we, we want it to be private for multiple reasons. We don't want to be surveilled by our governments or our, you know, other, other entities that can work, use that information for, um, nefarious means.

The next thing is bias and discrimination. And, you know, we train these models on lots and lots of data, and historically, some of that data may or not, may have been biased towards certain groups. So as we train these new models, we want to make sure that we evaluate the data so that it, it becomes a fair and ethical, um, framework.

So you wanna make sure that things like your loan applications at the bank are judged based on merit and not on any demographics. We don't want to discriminate against any groups. And there are tests that the machine learning community have put in place for many years that will help decide and inform whether or not the outputs of these generative a AI systems are fair and biased.

So along that same line that also we wanna make sure that we're inclusive, that we're, we're making sure that that, um, people have access to these new tools. Um, when the internet came about, um, there was a divi digital divide between the haves and the have nots. And, and over time, we've, we've made accessibility to internet via, um, you know, landlines and physical infrastructure or wireless through smartphones so that people throughout the world have access to that.

Um, we wanna make sure that not only do they have access to the internet, but also access to these generative AI models so that we have a level, um, playing field. And then finally, we have more on legal responsibility. And if you're a business leader, you want to make sure that legally you are adhering to all the guidelines for these new evolving laws.

The most, uh, if you do international business or if you're in Europe, the EU Act came out this year and they put up some, uh, legislation in place to make sure that our legal frameworks are responsibly managing, uh, generative AI as well as morality. We wanna make sure that these, um, systems and these are acting in a moral responsible way. And we do that by providing guidelines and sort of moral guidance for these systems.

The other thing I think that you should, should be aware of, especially as a business leader, is the amount of risk that could be, um, pro introduced from these ai. So if you look at ai, um, Aon, which is a global insurance and risk provider, um, they survey their customers and they're the ones that are underwriting the risk. And if they look at their, um, user base, they, they found that, um, their customers did not see the risk and ranked at 49th last year, while the person who has to, or the organization that has to underwrite these risks solid as 17th.

So I don't want to scare anyone that into thinking that there's a huge amount of risk. There's risk with everything. There's risk with what, um, how you measure your, um, treasury for your companies.

There's risk in going into new markets. You just have to understand that risk. And that's why we put on this, um, I give these kind of talks and put on these kind of events, is that we want to make sure that business leaders and, you know, regular citizens understand that while this is powerful and probably very good for society, there are risks and you want to take a measured approach to move forward.

And if I break down those risks, I look at the, the top three being inaccuracy. And this is from a McKinsey study. Um, and inaccuracy is often talked about in the media as hallucinations.

And what we mean by hallucinations is that the, uh, generative AI makes up or creates an FA, uh, a fact that is inaccurate. And the reason that happens is because we're really probability machines and they don't have enough background information to make a good, um, answer to your, to a query. So that's why you see so much, um, out there around data and licensing training data is that training data is one of the ways that you improve these, these systems.

Uh, the second one is cybersecurity. We're just increasing our PAC face. It happened when we, we became a more connected society, and it's always going to be us versus the bad guys, um, or gals.

And it's, it's just not, um, anything that's gonna slow down and it's, it's inherent in any new technology. And the third thing, and this goes back to the training data, is IP infringement. Are we using data that we are using responsibly and with the appropriate legal permissions to train our models?

So that leads us to the big question. If everyone is using artificial intelligence, how do you gain a competitive advantage over your competition? And that's really about the data.

And data's the new oil. It, um, provides, um, the ways for these, these machine learning systems to make better decisions, to create new products. You'll see this, whether it's in, uh, medical research or financial industries or a variety of other places, is if you have data and that data has the, is the dis distillation of how your business functions, you use that data to actually improve and learn and then draw insights for, for you to grow.

The other thing is we have lots of data that we are unable to access or not unable to access, but didn't have enough resources to, to access. So according to IDC, 90% of all data is unstructured. And a approximately, we only use 2060 to, uh, uh, about 60% of that one time.

And then it goes into data warehouses. We don't use it again. Um, the ability to have these automations that can draw insights from our data, um, that we don't get to is, is a huge advantage.

And so that's what I'm, um, one of the things that if you have data from your organization, historical sales, all sorts of, um, industry data, that's the kind of data that you're gonna use to inform your own enterprise GPTs or enterprise generative AI systems. And what you'll be doing is you'll use that data to improve your personalization for your customers, for your employees to make them more productive. You'll, you'll feed it into AI models, either ones that you build and host and tune or through the ones that are hosted for you.

You'll be able to do automation that you can be datadriven. A lot of companies like to say we're a data-driven, uh, company, but they're not an automatically data-driven company. As I mentioned earlier.

We're gonna, we're gonna get better insights because we'll have, you know, more, um, intelligence applied to that data. And then some of these companies, and then probably the one that's most, uh, uh, been in the news recently was Reddit. And they went public.

They've licensed their data to Google and allows them to Google to train their models, um, and that they're doing that with Google's permission. So that's great. And then generative AI is here to stay.

That's the takeaway I want you to have today. And it's gonna have massive impact. It's gonna, it's gonna affect all parts of our life.

And I think it's gonna be generally positive. You know, it's gonna improve our overall productivity. It's just like when we went from using hand tools to the tractor.

We have a, a country in the US that went from an 85% agrarian society to an industrial one, um, where only 9% of our workers are involved in food production. So that's, that's not a bad thing. I don't think many of us, uh, relish the fact of working in a field picking crops or hoeing the field using, you know, animals.

It's, it's a big, it's a big improvement. And as we adapt, we are gonna go from the same kind of things where things we did every day, they're no longer as menial and labor intensive, and we'll be able to use our minds more. And so my takeaway from everything today is that you should embrace a ai, or you most likely will be left behind.

And I don't mean that you have to wholesale live in ai, but if your business, your job, your, you know, your personal life is not being augmented by some of these advantages, you're not gonna have near the quality of life of those that take advantage of that. So I hope this was very helpful. You know, every, every week I publish insights like this in my newsletter, the artificially intelligent enterprise.

I hope you read it. I'd love to connect with you on LinkedIn or anywhere else. So thank you very much.

I.