AIリテラシーとは何か？ツールを使えること以上に必要な能力

The AI Today podcast, powered by the Project Management Institute, cuts through the AI height to deliver you proven strategies, real-world examples, and actionable insights that drive AI project success. Tune in to learn from PMI hosts and guest experts. The age of AI is transforming industries, creating a need for leaders who can leverage this innovative technology to drive real results. The University of Cincinnati's Online Master of Science in AI Management prepares professionals to lead AI initiatives across teams, projects, and entire organizations. Explore topics like AI governance, generative AI, machine learning design, and more, while customizing your degree with a stackable, career-focused graduate certificate. Build the skills needed to transform workflows and shape strategic business decisions with the University of Cincinnati. You can learn more at online.uc.edu slash AI. And we'd like to thank them for sponsoring today's episode.

Hello, and welcome to the AI Today podcast. I'm your host, Kathleen Mulch. And if this is your first time listening to AI Today, then welcome. And I'm so happy that you found us. And if not, then welcome back. AI Today has been going strong for 10 years now. We're on season 10, which is wild to think that we've never run out of things to say when it comes to AI Today. And if I'm being honest, I don't think I'm going to run out of things to say for quite some time. So if you haven't done so already, make sure to subscribe to AI Today so you get notified of all of our upcoming episodes. We have some really great interviews that are being lined up, as well as some additional topics that I plan to discuss. Sometimes I do interviews. I think that they're really wonderful and you get a perspective, right, of different folks. And other times I like to have a discussion where I, you know, talk on different topics and different topic areas, especially things that I've started to be seeing. So as many of you probably know, I travel quite a bit. That's wonderful, right?

I really get a global perspective and get to hear how AI is being applied and adopted and understood in different regions. And so that also helps me to shape some of the content for the podcast. For folks that have been longtime listeners, you know that the AI Today podcast and Cognolitica was acquired by PMI two years ago now. So we are officially a part of the PMI portfolio. You can find the podcast on PMI.org or your favorite podcast platform. And PMI is really leading the AI transformation of project management and the project management of AI transformation. And so that's really what a lot of the topics will be about, right? Because I want to make sure that we continue to lead in this space. So as I had mentioned, I get to travel a lot, which gives me really great global perspectives. And so I get to hear from practitioners. I also get to hear from thought leaders in the space. And I also get to hear from leaders who are working at various different organizations.

And one trend that has come up quite frequently now is I hear this term AI literacy. And many people are talking about it. Many people talk about how they want this AI literate workforce and that they're working towards getting an AI literate workforce. But they don't actually define what AI literacy is. And so I'm sure if you've been traveling and, you know, paying attention, going to in-person conferences, hearing panels online, you've probably heard this as well. And so I thought it was important to make sure that we have a podcast on this, right, so that I can talk about it. And as always, I love feedback from listeners. You can reach out to me on LinkedIn. I'm really accessible and easy to find. And shoot me a message and let me know what you think, right, if you've been hearing this AI literacy And if you've been hearing it defined as well, because what I hear far too often, right, is that almost every organization seems to be talking about AI literacy. Many leaders are talking about AI literacy. But like with many terms, when it comes to AI, it's hard to define.

So if I ask five different people to define AI literacy, I'm probably going to get five different answers. And some people may be talking about prompt engineering, right? To be AI literate means you have to be a good prompt engineer. Other people may be saying, well, you have to be proficient in tools. So whatever tool, AI tool or large language model of choice, it could be CoPilot, it could be ChatGPT, it could be Claude, you pick it, right? But as long as you're proficient, it means you're AI literate. And other people may say, well, it means you need to understand general concepts, right? So understand generative AI, understand agents, understand the latest models that are out there. And what I've seen is that none of these answers are necessarily wrong, but they're incomplete. And the danger is that organizations can start building training programs or just unleash tools and tell people to learn it on their own before they've actually defined the capability that they're actually trying to build, right? So they're saying, we want this AI literate workforce.

We're going to be giving you these different tools. Go out and learn them and become AI literate. So if somebody is really good at using ChatGPT or Claude or Copilot, does that really make them AI literate? And I say no, it doesn't, right? I said that it was not necessarily wrong, but it's an incomplete answer. And it makes them proficient in using a particular tool, which, of course, is valuable, but it's not the same thing as AI literacy. So it's important to understand here, tool fluency does not equal AI literacy. And I'm going to talk about this a little bit later, right, that there's different shelf lives when it comes to learning, when it comes to the AI era. And I've talked about this before, right, with my perishable, durable, and enduring skills that we talk about a lot at PMI. But I want to talk about tool fluency and AI literacy and how they are not the same thing, because it's this important distinction that using AI is not the same thing as using AI well, right?

And just because I understand a tool doesn't mean that I'm necessarily AI literate. So tool fluency is asking things like, can we use this interface or this tool of choice? Can you do things like write a good prompt or know how to use the LLM to help you write a good prompt? And can you use specific features? AI literacy is going to be digging one level deeper and asking these harder questions about, should I be using AI at all for this use case? What is AI actually doing, right? A lot of times, you know, hopefully we're changing the narrative on this, right? But a lot of people think that AI is magic, and it is not. You really need to understand how it works. Should you trust AI and should you trust the output, right? I always say trust but verify. So you can say okay this is the output that it gave me but let me use some of my critical thinking skills right and my judgment skills and my subject matter expertise in whatever industry that I working in to know if I can trust this output

And how do you validate that output, right? What risks may I be introducing by using AI? And what judgment still belongs with me as the human and the one who's ultimately going to be accountable for this? And then another really critical question to be asking is, did this actually improve the outcome, right? Well, you need to be measuring what the outcome is and what improvement looks like. So knowing how to get an answer from AI is not the same thing as knowing whether that answer is useful, trustworthy, responsible, and worth acting on. So if I think about, right, two different types of project managers. Project manager A is going to know dozens of advanced prompting techniques. They're going to use AI consistently on many things that they do. And they're also maybe going to know the latest models that come out from all of the different LLMs and, you know, what their names are, how to use them, all of that. But project manager B is going to know fewer tech sneaks, but have a better understanding of when AI adds value, how to check important outputs,

how to recognize when context is maybe incomplete or context is missing, how to protect sensitive information, right? We all have some form of sensitive information that we don't want out. And knows when human judgment needs to override an AI recommendation. So if I think about Project Manager A and Project Manager B, which one is actually more AI literate? That's a good question, right? And we have to be thinking about that. So I would say I think that's Project Manager B because maybe they don't know all the latest tools and features of every single model that's out there. but they're asking more critical questions. So what does AI literacy actually mean, right? Unfortunately, there's no commonly accepted definition of AI, so things just sometimes trickle down from there, right? If I can't understand what AI is, then what do I know what AI literacy is? And so I want to break it down into a four-part framework. Understand, apply, evaluate, and lead. So for understand, you do not need to have everybody on your team become an AI expert, right?

They don't all need to become data scientists or data engineers or AI engineers, right? I wouldn't actually recommend that everybody at your organization does that. But what you do need is people with enough understanding to understand what AI is good at, right, what it can and cannot do. when you should be using AI, when the outputs can be wrong, how outputs are influenced by data and also by context, and that different use cases carry different levels of risk. So that's also really important, too, because not every use case is going to have the same level of risk and maybe needs the same level of scrutiny and human oversight. Generating meeting minutes may need less level of scrutiny than something like, you know, loan decisions, for example, where it can really have a broad impact. So AI literacy doesn't also need to be identical for everyone, right?

I just said that we don't all need to become data engineers and data scientists. Well, we don't all need to have the same level of AI literacy. The CEO of your organization and maybe folks in the C-suite don't necessarily need to have the same depth of technical literacy as someone building an AI-enabled system. But that doesn't mean that the CEO and the C-suite don't need AI literacy. It just means that they need greater literacy around things like strategy and accountability, risk, investment decisions, and organizational impact. So that's the understanding level. Then we go to the apply level, right, when we're thinking about moving along. So can people recognize good AI use cases? And it's really important to understand it's not where can we put AI, it's where can AI meaningfully improve the outcome or workflow. And what I want to point out is that all of the workflows at your organization are pre-AI workflows. So what you shouldn't be doing is saying, here's a workflow that we currently have,

and let's figure out where we can bolt AI onto it so we can make it more efficient. Because if you've done that, you've probably realized it doesn't actually make it more efficient. It may just be creating different bottlenecks that get pushed farther downstream, or it's just making the process more complicated. You need to actually sit there and reimagine your workflow or your outcome. And that can be difficult. It's definitely not easy to do, but it's something that you need to be doing. Then we move to evaluate. And this is where things like judgment, right, these soft skills, we call them power skills at PMI, right, this is where judgment becomes critical. Because we should be making sure that, especially when we're thinking about AI literacy, right, can someone challenge an AI output? Can they maybe spot and identify where there's missing context and how to provide that additional context? Can they recognize questionable assumptions?

And can they verify important information? And then can they understand when additional expertise is required, right? And this also goes to that different level of risk, right? So not every use case carries the same level of risk. And then we want to go to leading, right? This is often missing from AI literacy discussions because AI changes workflows, it changes responsibilities, expectations, and sometimes it can even change and transform people's roles. So being AI literate isn't necessarily, it's, you know, about just understanding the tools, like I talked about. It increasingly means knowing how to help yourself and other people navigate that change responsibly. A lot of times people say AI and change management go hand in hand, and this is really where leading comes into play. How do you help others understand and navigate the change that comes about from AI. So I had talked a little bit earlier about the shelf life of AI skills.

And I think that it's really important to understand this because not all skills are created the same, right? When we talk about AI skills, not every skill we're teaching has the same shelf life. So when I think about perishable, durable, and enduring skills, perishable skills are going to be quick to learn and quick to decay, right? So they can be very valuable, but they're going to change rapidly. And unfortunately with AI, sometimes, you know, I can't even say in six months from now, it's gonna be the same because things are changing so often. But examples in perishable skills are things like today's prompting techniques or where a particular feature sits in an interface or what the latest models are on your LLM of choice or what you know configuration techniques need to be for a particular AI tool So it's not that we don't need to learn these things. We absolutely do need to learn these things. But learn them with the recognition that you'll probably need to relearn them fairly often.

And again, I can't say exactly how often, but know that it's going to be often enough. Then we have durable skills. And so these survive technology changes much better. These are things around frameworks and methodologies, right, like CPM AI methodology, which teaches you how to run and manage an AI project. Of course, what gets wrapped around it, some of those perishable skills can change. But the methodology itself hasn't changed since 2018 when it was first developed. Why? Because fundamentally how you run and manage an AI project hasn't changed. What also falls in durable skills is going to be things like how to identify valuable AI use cases. How do I evaluate AI outputs? Data literacy, which is also incredibly important and something that we don't talk about nearly as much. Risk assessments, AI governance, and also managing an AI implementation, right, especially if we're doing that with CPM AI. So a project professional who understands how to move from a business problem to an appropriate AI use case and then manage initiative responsibly has a capability that transfers across different AI tools, right?

So the durable skills are going to take a little bit longer to learn, but be more transferable. And then we have enduring skills. And this is really important, right, because these are around those soft skills like critical thinking, judgment, adaptability, communication, influence, change management and change leadership, and also curiosity. And, you know, all of these different soft skills, they aren't alternatives, right, to some of those perishable and enduring skills. but they're part of what makes you become good at working with AI, right? When you can apply your critical thinking skills and go, I don't think that that answer is correct, great. Challenge it, right? Or maybe if you use your communication skills, right, you can get even better at using some of these different things. And so this is where it's important. And this is also really important to understand because the better that AI gets, the more that things like judgment matters, right?

The more that these enduring skills really matter. And I think that this is important because as AI becomes more capable, sometimes people go, well, as AI becomes more capable, humans are going to matter less, right? As AI becomes more capable of doing things, then we don't need humans to do this. And in the news, there's been a lot of, especially, you know, in the past, probably six months to a year ago, people were talking about how AI is replacing jobs. And really what was happening is that organizations were right sizing or they were letting people go because they wanted to get, you know, instant cash flow so that they could invest in different AI tools and AI infrastructure. It's not really that as AI becomes more capable, humans matter less, right? I challenge that very hard, and I do not think that that is correct. What I would say is that as AI becomes more capable, the nature of human value changes. So if we think about what AI can do, right, especially these general purpose LLMs and these general purpose AI tools,

it's really good at doing things like generating and summarizing and synthesizing, analyzing, and recommending. But you still need a human in there to determine what problem you're solving in the first place, right? What business problem are you actually solving? You need a human to be answering that. Is the recommendation that the AI tool is giving you, is it credible? Is it accurate? Also, what's missing, right? Sometimes these tools don't have all of the context that you as a human do. So you need to say, I think that this is missing, and then challenge that, right? Also, there's going to be tradeoffs. So what tradeoffs are acceptable and what aren't? Who's going to be affected and impacted by some of these decisions? And what level of risk are we willing to accept, right? The tool is not, it should not be making that for you. It can provide those recommendations, right? It's good at doing things like providing recommendations, but at the end of the day, the human should be accountable for that.

And then how do you act as well, right? Should you act on certain things? Should you not? So when producing an answer becomes cheap, right, it's just going to be really easy. Deciding whether it's a good answer becomes incredibly valuable. And AI can accelerate good decisions, but guess what? It can also accelerate bad ones. And you've probably seen it accelerate bad decisions a lot more than it's accelerated good decisions. So that's also really important, too, right? Because we need to understand what data do we have? Is it acting on good data or bad data? And so how good are we at directing, evaluating, and acting on what AI does? That's really, really important. And so what I've seen now, I want to, you know, go back to what this organizational AI literacy trap looks like. Because first, we're talking about how we want an AI literate workforce. Okay, that's great. But then we need to be AI literate as an organization, right?

So now we need to move from individual AI literacy, which organizations still aren't, and leaders, right, aren't really defining super well. So now I've gone through a few different ways that you can think about that, right? So how do you understand, apply, evaluate, and lead? How do you understand that the shelf life of AI skills are going to be different? So we have perishable, durable, and enduring skills. and why those enduring skills, right, especially as AI continues to get better, are more critical than ever. But what does this AI literacy look like in the organization? Because now, at some point, we're going to have to move from individual AI literacy to organizational AI literacy. And this goes deeper, right? This needs to go a little bit deeper than what we've been talking about. So organizations often respond to AI literacy with things like tool training. Maybe they'll give you live in-person training. Maybe they'll tell you to go watch some videos. They're definitely going to give you licenses, right?

And maybe they're giving you access to too many AI tools or you just don't have the bandwidth to actually learn all of them, right? And some organizations are giving you usage targets. So they'll have leaderboards and they'll talk about how, you know, different folks are using AI. and they'll keep this very open and public, what that does is that encourage shallow use of AI, right? It encourages a lot of shallow use of AI so that you can reach these usage targets. And some of this is okay but it really you know high AI usage doesn actually equal high AI maturity You can have that high but shallow usage So we really want to be asking, are we measuring AI literacy by how many people are using AI or how well people are using it? And this is a shift that we need to make sure that we're making. So as individuals, right, I'm definitely encouraging you to be AI literate. Keep listening to AI Today podcast.

That's one great way to do it. PMI is a lot of different resources that are out there, right? We have eight e-learning courses that are all free for members. Some are free for non-members as well or carry a small fee so that you can take that. There's also a lot out there right on YouTube, on LinkedIn Learning. That's how you can keep your perishable skills fresh. So learn today's tools enough to be productive. I want you to experiment, try new capabilities, ask questions, right? At PMI, we also have chapters that are global, so I encourage you to reach out to them, reach out to peers so that you can learn from others. But understand that these perishable skills are going to have a shelf life, and that's okay. That's why you have to constantly be learning, right? I talk a lot in my podcast about this idea of a growth mindset where you need to be this lifelong learner. And then invest intentionally in durable skills, too. So one thing that is so critically important is use case evaluation, right? Because people can have many ideas of how to use AI.

And I'm sure that if you ask to get a list of use cases, you'll probably very easily get dozens, if not hundreds, of different use cases. But now you need to go in there and evaluate, is this actually a good use case for AI? What is going to be the return on that investment for us to use this, right? Because if it's something that you're doing daily, maybe that's a great use case. If it's something that's only run quarterly or once a year, maybe that's not a great use case for AI, right? It might just be quicker, cheaper, easier to do that manually, or we are going to push that down in the priority of this. And then also durable skills are going to be things like learning CPM AI, right, which is a methodology for running and managing AI projects. And it's also our fastest growing certification at PMI. And then I say really invest heavily in these enduring skills, right? It's going to take a lot longer to learn, but it's a lifelong learning journey.

And so these are things like judgment, critical thinking, adaptability, communication, influence, change management, because you need to constantly, constantly be practicing these skills. There was a study that came out from Microsoft and Carnegie Mellon about a year or so ago, and it said that the more we use AI, the more our critical thinking skills will go down. And once they go down, they're really hard, if not impossible, to get back. So I encourage you to continually applying them, right? Challenge assumptions. Use your judgment. Use your critical thinking skills. And use your communication skills, your influence skills, right? Use what makes you human, especially the more you're using these AI tools. So you really should be asking, is our AI learning strategy preparing people to use today's AI tools, or is it preparing them to continue making good decisions as these tools change and evolve over time?

So I want to leave you with this. AI literacy is not about knowing every single thing that you can when it comes to AI. In fact, if you do that, I always say don't go tool chasing, right? Don't go chasing every new announcement that comes out. It's going to make you go crazy because the announcements are happening way too often that all you would do is be focusing on that, and that's really just focusing on those perishable skills. You need to make sure you have this well-rounded learning experience, right, of perishable, durable, and enduring skills. So don't go chasing every single AI tool and knowing everything there is to know about AI. Why? Because that's just part of the equation. Instead, you really should be understanding AI well enough to use it appropriately, evaluate its outputs critically, manage its risks responsibly, and adapt how work gets done. So learn the tools well enough to participate, to have meaningful conversations, and to be able to use them in an effective way.

Build durable capabilities so that you can adapt, right, and that you can understand use cases, how to evaluate use cases, how to run and manage AI projects. And then strengthen your enduring skills so that you can lead, right, and you can confidently use these tools, especially as they change. So the goal shouldn't be to create a workforce that can keep up with every new tool. Hopefully by now I've harped on that enough that you understand, yes, you need to use tools, but it's not just about that because that's probably impossible, like I said, really, especially learning every single one. The goal should be to create people and organizations capable of continuing to learn, adapt, question, and exercise good judgment no matter what the next tool is that you will be using. And so, you know, I really want to leave you with that. I would love to continue this discussion, right? AI literacy is a topic that I hope to continue to revisit. So please reach out to me.

Like I said, I'm available on LinkedIn. You can comment on my posts, especially the ones that I post about this podcast. And let's continue this discussion because I'd love to hear your definition of AI literacy. I'd love to see how your organization is bringing AI literacy into the conversation and how this continues to evolve over the next couple months. So if you haven't done so already, subscribe to AI Today, and also please make sure to rate the podcast and always reach out to me. The age of AI is transforming industries, creating a need for leaders who can leverage this innovative technology to drive real results. The University of Cincinnati's Online Master of Science in AI Management prepares professionals to lead AI initiatives across teams, projects, and entire organizations. Explore topics like AI governance, generative AI, machine learning design, and more, while customizing your degree with a stackable, career-focused graduate certificate. Build the skills needed to transform workflows and shape strategic business decisions with

the University of Cincinnati. You can learn more at online.uc.edu slash AI. And we'd like to thank them for sponsoring today's episode. Thanks for listening to this episode of AI Today. please make sure to rate this podcast on Apple Podcasts, Spotify, or your favorite podcast platform. And if you want to lead both the AI transformation of project management and the project management of AI transformations, check out resources from the Project Management Institute, including the CPMAI certification, our various e-learning courses, and the PMI by AI blog. Until you join us next time, keep innovating. Thank you.