Transcript
Speaker: Made on Zencastr. Because Zencastr is the all-in-one podcasting platform that really does make every stage of the podcast production and distribution process so easy.
Speaker: All the details are in the description. Hello and welcome to The Independent Minds, a series of conversations between Abysseedah and people who think outside the box about how work works with the aim of creating better workplace experiences for everyone.
Speaker: I am your host Michael Millward, the managing director of Abbasida. In this episode of the Independent Minds I am talking to Vin Mitty, the author of the AI Decision Map.
Speaker: Vin is based in Oklahoma, which is not somewhere that I have ever visited. When I do get the opportunity to visit Oklahoma, I will make all of my travel arrangements with the Ultimate Travel Club.
Speaker: That is because as a member of the Ultimate Travel Club, I get access to trade prices on flights, hotels, trains, holidays, well, all sorts of travel related purchases.
Speaker: I have added a link with a built-in discount to the description so that you can also become a member of the Ultimate Travel Club and just like me, travel at trade prices.
Speaker: Now that I have paid some bills, it is time to make an episode of The Independent Minds that will be well worth listening to, liking, downloading and subscribing to.
Speaker: And I am sure it will also be good enough to share with your friends, family and work colleagues as well. As with every episode of The Independent Minds, we will not be telling you what to think, but we are hoping to make you think.
Speaker: Now, hello Vin. Hi Michael. are you today? I'm good. How are you? Not too bad at all. Thank you very much. Please could we start by you telling us a little bit about your history. Who is Vin Mitty? Yeah, absolutely. i was born and brought up in India. I started off my career as as an entrepreneur building websites and ERP solutions for small businesses and NGOs.
Speaker: Since then, I've been a data and AI leader for about two decades. Like I said, experience in technology, machine learning and AI systems. I've worked and consulted with startups and Fortune 500s.
Speaker: I've helped companies move from Excel spreadsheets to machine learning predictions to AI chatbots that help them grow their businesses and retain their customers.
Speaker: Today, I lead data science and AI initiatives at a company called LegalShield in Oklahoma. My team builds data and AI systems that influence marketing, operations, and product strategy.
Speaker: Alongside my work, I also completed my PhD research on why organizations struggle to adopt data and AI technologies. That combination of practitioner experience and research has really helped me shape different perspective on ai when you say different perspectives on ai one of the things that i'm thinking is that there may be seven billion people in the world and let's assume that a billion of them have knowledge of or some sort of link connection usage of something that is ai but every single one of them would have a completely different
Speaker: view of what AI is and how it should be used to fulfill its potential. Absolutely. Yeah. I mean, I think AI is so vast, the applications are vast, so all of us have different use cases. Absolutely.
Speaker: Yes. I totally get what you mean, I think, by totally different use cases and there doesn't seem to be a standard definition of what AI is. And I suppose That's part of the reason why we need a book about the the AI decision map.
Speaker: Yeah, absolutely. And like you said, there's so many different definitions of AI, you know, and also I did want to point out that AI is not new per se. We've been talking about and thinking about artificial intelligence since 1955. when John McCarthy, the scientist, computer scientist, na coined the term AI or artificial intelligence. So the technology is being worked on or being thought of for for decades now. Speaking of the book itself, The AI Decision Map, the book came from a question I kept asking throughout my career.
Speaker: Why do smart organizations make bad decisions about technology? I kept seeing the same patterns, right? So in the early web era, the mobile era, and now with AI, there is a lot of hype about a new technology and it's positioned as the savior, the answer to all of our problems. We tend to use it as a hammer to solve everything.
Speaker: And then we're all disappointed. With AI specifically, a company gets super excited about AI, launches, pilots, builds dashboards and whatnot, and six months later, later no one's using these shiny, tall.
Speaker: That is so true. But I think it goes beyond ai In many cases, organizations invest in technology but don't actually train people how to use the technology.
Speaker: So a classic example would be that people use Word like a typewriter, but it's capable of so much more. Because nobody's really been on a training program in how to use the computer that they've been given.
Speaker: We spend a thousand pounds on a laptop for an executive and they've got it in the the nice bag and they carry it around. But it's not being used. People are very rarely trained in how to use it.
Speaker: In my experience, at least, the training that people do get, like you say, isn't connected to how they need to use it in their jobs. A hundred percent. So yeah, in, in my, my research and also in my career, I've seen that more often than not, it's not the technology that's the issue.
Speaker: It's other things around it, like. The human being. The human beings. Exactly. Exactly. The human being and I suppose the original decisions and the context in which those decisions are made.
Speaker: But in terms of the research that you've done. Before you talk about that, I'd like to a decision map. It sounds great and I like the idea of it, but I really want to clarify what you mean by a decision map.
Speaker: Yeah, so every business leader is under pressure to do something with AI, but we see 95% of AI projects fail. And like I said, I've been working in the area for for decades and also my research is in that. So I've gathered my thoughts into frameworks so leaders can go beyond just the hype and you know the bubble and deliver real value with AI. So this book gives frameworks to think about implementing AI at your business, at your company, at your Fortune 500. When you say Fortune 500, is it specifically targeted at that group or could it be applicable to any type of organization?
Speaker: it's it's ah it's It's from startups to small organizations, to NGOs, to any anybody can use it. In the book, I have almost 80 pages of examples to show how my ah my frameworks can be applied to different industries and different sizes of organizations, from healthcare to marketing to manufacturing, et cetera. And all of these frameworks are based on the research that you've been carrying out.
Speaker: and also your personal experience. So tell us a little bit about the research that you carried out. what What sort of things were you looking into and what did you find? Yeah, ah my PhD research focused on why organizations adopt or resist data and predictive technologies. What the research showed was very clearly is that the barriers to AI adoption are rarely technical, like you said. They're human and organizational.
Speaker: people resist systems they don't trust, they worry about being replaced, or leadership hasn't clearly connected the technology to real life outcomes. So I think at the end of the day, the takeaway is AI adoption isn't really a technology problem. It's more of a leadership and more so a trust problem.
Speaker: I get that. I do. And I'm remembering the things that are being said by some very prominent figures about how AI is going to replace a lot of the mundane jobs that people have to do. It won't be done by a human being, it'll be done by AI.
Speaker: think that fosters the belief that AI is going to do away with lots of jobs. But what I'm seeing is that, and what I've seen other research demonstrating is that far from doing away with jobs, it can actually expand employment opportunities as well if it's used in an appropriate way.
Speaker: A hundred percent. So I think AI is kind of forcing all of the white collar jobs to evolve. Right. So it's it's helping people elevate everybody into strategists, you know, from from an analyst can become a from ah an analyst can move from being a report runner to a business strategist, for example, or a marketing person, instead of creating hundreds of emails, that they can have AI do the first draft and think about really how they want their brand to be communicated with their customers. Instead of focus on focusing on grammar and sentence structures, you can think about brand and how you, the the feel of how it comes comes up across. What you're actually saying there is the feel and how something comes across.
Speaker: is one of things with my HR professionals hat on is where I think AI enables us to do more of, which is like AI do the technical side of things, the run of the mill type things.
Speaker: And then we can do more of the emotional side, which AI can't do at the moment. We can do more of the human to human interactions because AI is dealing with all the other things in the background.
Speaker: Absolutely. Absolutely. You mentioned that the research that you've you've done has led to the result, you've been able to create 88 pages, was it? Yes, yes. Of frameworks.
Speaker: So tell me a bit about the frameworks. We can't go through all 88, but which of them do you think would be the most useful for, say, a small employer?
Speaker: Yeah, no, um yeah it's it's ah primarily two primary frameworks. It's called the AI decision map and the AI value review. Let's focus on AI decision map, for example. So it talks about Where AI lives. So I'm asking the you know readers to think about three questions, right? Where to play, what to solve, and how to think in the context of AI. So you're asking where the AI lives. Is it internal or external? Depending on that, you have different risk propositions.
Speaker: And then what to solve. What does it do? is Does it automate? Do you want it to augment? or do you want it to autonomously do something? And then how to think.
Speaker: Do you want it to predict? Do you want it to generate something? Or do you want to it to just go do things for you? So depending on these, there are different risk factors. There are different ah things to think about, which I talk about in the book. So I think the 88 pages are really applications of these frameworks in different industries, right? So how would you think about it when if you're writing emails? of How do you think about if you're applying AI in the HR world, or manufacturing world, and so on?
Speaker: So you start off with an overarching type framework and then break that down into different industries, different professions, different scenarios to demonstrate how AI could be applied.
Speaker: And I sort emphasize the word could be applied in those types of environments. 100%, 100%. Yeah. and And then we also talk about what we just talked about, right? So it's not just the technology, it's all about how you approach the technology. you is your Are your employees or is your company ready and trust the technology?
Speaker: Are they embracing it or are they in fear of losing their jobs? Yeah. I get the feeling your research will show that Many organisations are not actually quite ready to answer those questions.
Speaker: They don't know themselves well enough in order to be able to answer those questions, well, in an appropriate way, in an honest way. We can very much get wound up in new technology new ideas and almost say, we must have it as almost a a fear of missing out type approach to it Exactly, exactly. I've sat in you know executive meetings or boardrooms where people come in and say, hey, let's just use AI, right? But we have to really think about, start from what problems are we solving and is AI the right solution for it?
Speaker: Yes. Sort of thinking that The same problem, the same challenge, might exist across numerous different organizations in the same industry and other industries.
Speaker: But unless you know yourself as an organization very well, you won't be able to identify whether AI is the right solution for you or you should be doing something else.
Speaker: Whereas it might be the right solution for another organization, but it won't be the right solution for you. how can an organization identify, work out what where they are in that AI adaption roadmap?
Speaker: Yeah, no, absolutely. I think one of the main questions is to kind of and understand fundamentally what AI is good at and what it is not, right? So if you want AI is basically a probabilistic model, right? so it it So it predicts what the next word is or what the next pixel should be, right? So it it is thinking probabilistically. So if you ask it the same question
Speaker: 10 times, it'll give you slightly different answers. So having this context in mind is really helpful, right? So if you are asking it to do math or accounting workflows, maybe ai is not the the right solution. Maybe software is the right solution because you need the deterministic. ah You need one plus one should always be two.
Speaker: But with AI, one plus one could be three, one plus one could be four or one, depending on probabilistically what it thinks the next...
Speaker: answer should be. So AI solves a lot of problems. It can do a lot of creative work. It could do, it it is really, really good at summarizing documents or reading through a lot of things and gathering insights from it, but it's not really spot on with math problems per se, accounting problems. So understanding that AI is probabilistic and not deterministic could be the first step to ah seeing whether AI is the right tool for your problem.
Speaker: Yes. So in the problems that I see organizations having with a eye is the expectation, if not sometimes the belief that AI is going to be by definition correct.
Speaker: And your job is to disprove the AI. Whereas what you're saying there is that if you want a definitive, definite, no-quibble type answer, then you want you need software in order to create that. Two plus two is always going to equal four.
Speaker: But because of the probabilistic, the probability type model that AI uses, which you can see when you use Word and you're typing in a word and as you type extra letters the auto-completion part of Word changes the word that it is going to suggest that you want to use or even the expression that you want to use.
Speaker: That I think is that's a real eye-opener then about what AI is and what AI isn't. It's just simply that that acceptance that ai works on a balance of probabilities and software works on this determined it's a definite fact in ai in in software whereas ai works on the balance of probabilities 100 you put it in a great succinct way yeah only because you'd explained it
Speaker: But that is that is part of the challenge with something like AI. It's the human element that is assuming that AI is infallible in lots of ways.
Speaker: And yet that assumption that it's infallible is probably one the reasons why AI adoption projects fail. hundred percent. So I think that there's a lot of expectation on AI that it it's going to solve everything.
Speaker: And there's also a lot of frustrations when it doesn't and it it errors out because it's applied to the wrong type of problem. It's not the AI.
Speaker: It's that we're asking it to look at the wrong type of problems. Exactly. That makes an awful lot of sense. So the danger is, I suppose, that people use AI for things that AI is not the best solution for.
Speaker: It doesn't work. They lose confidence in their use of AI and confident they lose confidence in AI itself. What are the quick, easy, simple, straightforward ways for an organization to use AI in a way that enables AI to demonstrate its full potential.
Speaker: Yeah, there's there's a lot of examples of where AI could be super helpful. Again, it anything that you, wear where you need ai to synthesize a lot of information, AI is really good.
Speaker: Any places where you need the first draft of a creative solution or brainstorming, it's it's really good for for that. And now it's really, really good getting good at coding. So if you want to build a website or build an app, Cloud can do a wonderful job of creating those softwares for you. But again, it's not that you fire all of your software engineers, right? Because yeah you need, again, like we said talked about, it's not perfect.
Speaker: So you need humans to test it, make sure that it's working as intended. yeah A lot of our jobs will ah will evolve into becoming more of the editor editors and quality control people versus development itself. Right.
Speaker: So if we're going to become the editors, the checkers of what the work that the AI has done, how do you think we are going to develop the knowledge, the skills, the experience to become that role?
Speaker: Whereas before AI, You would start off at the bottom of the career map and you would learn all sorts of different things and build your knowledge. Now, to a certain extent, it seems like you don't need to you won't need those, that knowledge, you won't be able to get that because AI is doing those early career type jobs.
Speaker: Well, how do you build the skills, knowledge and expertise to be the editor if you haven't got those lower level jobs? Yeah, I think of it as a natural evolution when a new technology comes in, right? For for my generation and kind of a few in the future, i mean, future generations as well, I don't know how to read a map. Once Google came in and the search engines came in,
Speaker: there was less incentive to remember facts anymore. It's not that I don't remember anything or I cannot ever read maps. So it's it's a skill that's becoming less useful because I can always look it up.
Speaker: right So similarly, I think In the early stage careers, i think we need young people to provide more value using AI technology. So we need young people to become experts at AI and provide more value than you know, the young people or early career folks were providing a couple years ago. So so I think we can all become 2x, 3x value providers for our employers than what we were we would with without AI. So I told you I started off my career as an entrepreneur
Speaker: i think one thing i learned there is the value and the the money you make or the value you get is almost directly proportional to the value you provide to others so think about what value can you bring this you have a really good tool that can help you out so think about what your expectation what the expectation is from you and deliver value at a higher level because it's getting easier to do that.
Speaker: Yes. I suppose essentially, as you're saying, AI will not do away with every job. We will become more adept, more skilled at using the information that AI generates for us.
Speaker: We will become probably more discerning about that information and probably be using different forms of AI, asking the same question and then triangulating the information that we receive back from the different versions, using them as sources of information rather than just something that is infallible. Oh, 100%. The way to build a career where AI is part of the working environment is to work with the AI
Speaker: and use it as a way of building your skills so your skills, your knowledge and your expertise so you become an expert in using the a i rather than an expert in what the AI is doing or as well as an expert in what the AI is doing. 100%. I think it's it's a combination of both. Yes.
Speaker: Yeah. It's really interesting. And I cannot wait for the book to be published. It sounds like it's going to be a very interesting read. And it's only a matter of days before it is published. We will include link to the the advance order page for the book in the description. But for today, then, it has been fascinating.
Speaker: Thank you very much. Really do appreciate your time. Thank you so much. It was really fun. Thank you. I am Michael Millward, the Managing Director of Apocida, and I have been having a conversation with the independent mind Vin Mitty, the author of the AI Decision Map.
Speaker: You can find out more information about both of us by using the links in the description. AI is playing a big role in modern healthcare, care but that doesn't take away our responsibility for maintaining our good health.
Speaker: And the best ways to stay healthy is to know the risks early. That is why we recommend the health tests available from York Test, especially the annual health test.
Speaker: The annual health test from York Test provides an assessment of 39 different health markers, including risks of chronic illness like cholesterol levels and diabetes, vitamin levels, organ function, as well as a full blood count.
Speaker: The annual health test is conducted by an experienced phlebotomist who will complete a full blood draw at your home or workplace. Hospital standard tests are carried out in a UK AS s accredited and CQC compliant laboratory.
Speaker: You can access your easy to understand results and guidance to help you make effective lifestyle changes anytime via your secure Personal Wellness Hub account. There is a link and as you would expect a discount code in the description.
Speaker: I am sure you will have enjoyed listening to this episode of The Independent Minds as much as Vin and I have enjoyed making it. So please give it a like and download it so you can listen anytime, anywhere.
Speaker: To make sure you don't miss out on future episodes, please subscribe. You'll probably also want to share the link with your family, friends and work colleagues as well.
Speaker: Remember, the aim of all the podcasts produced by Abbasida is not to tell you what to think, but we do hope to have made you think. Until the next episode of The Independent Minds, thank you for listening and goodbye.
