Start With AI Literacy: Why Most Leaders Are Missing the Point

**Alexis Waringa** (0:03)
Welcome to Gartner ThinkCast, I'm Alexis Waringa. According to a recent Gartner survey, nearly half of CEOs think their AI investment is break even at best, and only 5% even call it a real win. But why is that? Today, we're exploring why the bottleneck was never the technology. It's that many of us were never taught how to actually think alongside it, question it, or catch it when it's wrong.
To help go through what AI literacy actually looks like in practice and how you build it at scale, we're joined by Gartner Distinguished Vice President Analyst, Mandi Bishop. Mandi, welcome in.

**Mandi Bishop** (0:35)
Thank you so much for having me, Alexis. I'm really thrilled to have an opportunity to talk about this topic. I'm so passionate about it.

**Alexis Waringa** (0:41)
Yeah, I'm excited to talk with you. So right off the bat, you have a blunt message for CIOs and AI leaders, which is you're missing the point. What exactly is that point that they're missing?

**Mandi Bishop** (0:53)
Yes, absolutely. So I hear all the time that we're not getting value from our tools, right? We've invested all this money in co-pilot and all these licenses, and we're scaling these tools across the enterprise, and people are using it for a week, and then they're not getting anything from it, or adopting these models, or making these things.
And that's great, and you need to be doing that, but you are not addressing the challenge of AI enablement for your organization, right? You're not helping everyone learn the critical skills that they need to be able to safely and effectively interact with these tools at scale, no matter what tool it is, right? No matter what technology you're using. That's not the differentiator. The people are and their ability to leverage AI, right? To actually drive that value.

**Alexis Waringa** (1:43)
And why is maybe the idea to buy more tools usually the wrong next move?

**Mandi Bishop** (1:48)
Doesn't matter how many tools you buy, right? So I hate to cook.
It doesn't matter how many cookbooks I have. It doesn't matter how many fun gadgets I buy. Two years ago, I bought an air fryer because everybody told me I needed one, and it looks like it's amazing, and I know it can do awesome things. That thing is still in its box somewhere in my house. I can't even find it. My husband has hidden it because he gets so frustrated when I buy these things, right? It just because we have the tools that can do amazing things, doesn't mean that we have the understanding or the desire, right? The motivation to really use them effectively.

**Alexis Waringa** (2:22)
So where does that gap actually show up in the day to day? What does a lack of AI literacy look like in practice?

**Mandi Bishop** (2:29)
Sure. So it shows up in a lot of different ways.
As we use AI more and more, right? It's not just about, from a literacy perspective, knowing that we can use the tools. It's not about understanding, OK, I've got CoPilot. I can use CoPilot to summarize email. Simple thing.
If you use CoPilot routinely to summarize email, and you don't, you get good enough answers, you don't think about, is that the right answer, right? You start to really trust it too much. So you develop automation bias, and that can have really significant downstream consequences for the organization. And being able to identify and really help users think through, how do I validate? How do I continue to make sure that in my use of AI, it is reliable, right? And that there are the consequences of the decisions that I'm allowing it to make, this type of automation bias, are something that I can prevent upfront. Yeah, but so it can show up downstream. I work a lot with health care, so you think about if I am a case worker for someone who's been discharged from the hospital, and I have tools now that can summarize the information from that hospital encounter, I review that summarization, and the AI tools give me that summarization, then you can click through and it gets pushed downstream to their health care doctor, right? So it gets pushed downstream to be used in clinical practice. Well, if I think the summary is genuinely okay, and I stop really reviewing it, and I stop actually making sure it's correct and that it's complete, then the quality degrades and it could actually end up harming someone downstream. So whether it's a human that it harms or whether it's the organization that it harms, there's very real consequences to not getting this right.

**Alexis Waringa** (4:30)