Ep208: Using AI to help take a load from our busyness backpack (First Friday)

John Barrett talks about how we can use Artificial Intelligence to help when we are feeling overwhelmed and our virtual backpack is overloaded.  Among his suggestions are:

 Think of AI as a super-smart, literal and naive intern, who can’t read your mind but who has unlimited time and access to information. And who will provide you exactly what you ask for with high probability. Use AI to help you use AI

• • What is an effective prompt to …..?
  • What questions do you have for me to better answer my request?
• Provide the AI the parameters for exactly what you want (don’t worry about the HOW)
• E.g.
  • Answer as a thought leader who has deep understanding and practical experience
  • Your audience is junior employees who are smart and inexperienced in the workplace
  • Define the key thinking and provide 3 practical actions they can take to be better team members
  • Provide references and a list of further resources for deeper understanding
• Adopt a 3 iteration approach
• Give feedback and/or adjust request and ask for another suggestion

He also mentions a couple of resources...

• https://www.therundown.ai/subscribe
•  Use the Prompt Engineer GPT by The AI Report to 10x your prompts.

Transcript

Humanergy (02:22)

First of all, for those of you who've been to the previous sessions, we've been talking about this idea of managing your workload. And whether it was the idea of a backpack or, you know, a bucket, we've been using this idea of a physical metaphor to get a key point across, which is, you know, reality is brutal, right? Our capacity is fixed and, you know, whatever doesn't fit in, it doesn't fit in, in terms of workload.

Now, what we find is that's quite often hidden to people. You know, we experience the symptoms downstream, but at any given time when we say yes to something, we don't have a physical signal, if you like, or a strong signal that says, “Hey, this is too much.” So the thing about thinking of it in a physical way—whether it's a backpack, your suitcase when you're going to go on a business trip, or a bucket—is that you soon realize that every time you add something, something comes out. Right? And if you take something out, then you have the potential for something in.

So I just wanted to sort of capture that idea as a starting point. Now, for those of you here last month, Lance talked about this idea of “get a larger bucket.” And that's what we want to explore today in terms of AI as a partner—partnering with AI. So in a sense, expanding your bucket by having a partner with AI.

And just to sort of build that idea in, I just wanted to share with you—there we go—a quick video. It's one minute long. Just a quick background: Mount Tai, which is where this footage comes from, is a famous place to visit. It's a beautiful place where people pilgrimage. And the way you generally get up to the top is a beautiful pathway that is 7,200 steps tall and takes anywhere between three and six hours to get to the top.

Now, you can cheat. You can go around the back and get a bus and then get a cable car. But for most people, the experience is to go up. And there are people who live all up and down this pathway. And so, you know, a lot of people—the way that things get up and down in terms of goods and services and, as you can see, trash—is it has to get literally carried.

So here we have an idea of how we can expand our capacity.

Okay. So, you know, carrying a load up a steep mountain is not something that we're going to be doing anytime soon. But the idea being is that, you know, if we want to expand our capacity, one way of doing it is to get some help. And in fact, a robot is where a physical device and AI actually come together. And in fact, it's the AI that makes the magic in this case.

But a couple of things—three ideas for the balance of our session that I have here:

1. How can we engage AI to do more—right? Carry more weight, carry more load, 40kg, 80kg, 120kg?
2. How can we engage AI to be more? Meaning, if I'm not loaded down, sweating, carrying a huge weight, maybe I can have a different experience of the mountain—right? And if I'm not loaded down and I have the bandwidth, what things can I see? What things can I do? What things can I be?
3. And then lastly, the last thing I want to talk about quickly is: how do we grow wiser? Because the idea of partnering is very much about a learning experience—of both what AI is capable of and what it can offer in doing more for us, and then what is the impact on us of doing less?

So let's jump in.

First of all, as you're probably aware, this AI thing is not going away. It's like an avalanche. If anything, it's growing in size and volume. And every day I look, something new is coming along.

The other thing—and research shows here, and this is just one of many, but this is from McKinsey—is that up to 30% estimated of current work hours may be potentially replaced across the board through this AI piece. And as you can see from this graph, it's not physical work per se—that's pretty much stable. It's data management work—so collecting, processing data—but then stuff that we would typically say, “Well, isn't that what people do?” Interacting with stakeholders, managing, developing people, applying expertise. And you can sort of see the estimates are not just 10%—it's 200%, it's 300%. I would anticipate actually this is an underestimate.

And one last piece before we move on. This is an early piece of study. Harvard study—people looked at Procter & Gamble, 700 professionals, and they looked at scenarios where you had teams—so people and people working together—but then teams with AI. And what they found is that teams with AI were the most productive combinations thereof. Not only getting more done, but also producing better solutions. And as it identifies here, it's not just a quantity of work, it's also a quality of work. And what I find is fascinating is that the teams that were engaged with AI actually had a different experience—not that they were less satisfied. In fact, they were more satisfied and had more positive emotions overall.

So let's explore this idea of “do more.”

I actually used AI to generate this. And so here's an example of how we can do more with less. If I initially was going to put this list together with my own notes, and then I thought, well, let's see what AI can do. So you can see my prompt at the top:

“Generate a list of the top 10 easy and best uses of AI to help a professional in everyday productivity.”

Right? So within about 15 seconds, I had this list. I was then able to spend a couple of minutes looking through it. And what I was able to do very quickly was to say, “You know what? I can relate to this. This matches my personal experience and others' experience, step by step.” So I was able to validate it quickly.

So I'm going to just invite you to cast your eyes down here just to spark your own thoughts about possible uses. And also, you probably see some things that you're already doing.

Okay. Then for fun, I decided, well, let's just see what happens when I just ask for another 10. And so again, about 15 seconds later, I had another list. And again, I spent another couple of minutes looking through my notes, looking through my own experience, and realizing that this really was an excellent summary in order to build on.

And again, I share it with you, encourage you to have a look down through and see where you potentially have some opportunity.

So I wanted to give an example very quickly of how we're looking at exploring something. We do 360s for leaders, and it is a fairly complex and sizeable piece of information. People get feedback from, you know, 10 to 20 people on 75 questions and then are invited to provide written comment—multiple times throughout. This generates a report of about 40 pages of data, as well as up to 20 pages of written comment. So it's sizeable.

Typically, it would take one of us at least two hours to look through it and to digest it and to provide a key point summary and insights to somebody. And we also have people do it themselves—typically that takes them about four hours.

So with a little bit of experimentation, what we found is we could literally take this PDF and, without any education, any training, or anything like that, literally put the PDF directly into the AI platform and it would provide a fairly detailed and insightful analysis.

Now, we just don't use it like that. What we do is we then take that—and we're currently in the process of comparing it with the human form analysis, of which we've got now hundreds—and the AI, and seeing what we can learn about from that in terms of partnering.

So this opportunity for being more—sorry, doing more.

Now, being more is not just, “Hey, I used to do this and now I don’t.” Right? The thing is, AI can do things that you can do. But for those of you who've played with it, I think you sort of see there's a possibility that AI can do things that you can't do. And in doing things that you can do, it releases you to do things that you don't have time for.

So between doing things that you don't have time for that have more value, plus tapping into the unique qualities of what AI can bring, and then adding to that your high-end expertise, then we can—in my mind—we can be more.

So I just want to share with you just some of the things, just to sort of stretch your mind, that AI can currently do. And again, I'm not going to go through all of them, but I just wanted to put this top 10 in front of you.

And as you're reading that, a couple of other additional ones—for example, AI currently tests out in the top 10% on standardized law exams. AI currently tests out as being outstanding and actually best at doing some high-level medical analysis.

So my point is that the limit in my experience of how much we can use AI is generally when we look in the mirror. The limit is our imagination of what is possible.

Grow Wiser

And lastly, grow wiser. And there are a couple of things that we need to be aware of.

First of all, AI—even as it gets “smarter”—only emulates intelligence, insight, and understanding. It doesn’t actually have those capabilities inherently. And as you're probably aware, it does have biases. It is also subject to illusions, etc.

Now what I would notice is—so are we. So it's not like, “Hey, AI is flawed, you can't trust it.” I mean, I think if we look at what we do too, we also have inherent limitations.

The second thing I want to highlight is that this idea of unearned knowledge is both seductive and dangerous. There is something to be said for doing this with equity yourself. I’ll give you an example.

Back to the 360. We insist that people do their own analysis of their 360, and it takes—like I said—up to four hours. And the reason is that that experience is transformative. If we just gave people, “Hey, here’s the top key point analysis,” it would not have the same impact.

And there are lots of examples of that—you can see a summary, think, “Oh, I understand that.” But if you haven’t earned it in some way—either the sweat equity of doing the base research or doing it yourself, etc.—then your understanding is paper-thin to some degree. And you have to be aware of that.

Third point is that there is a confidentiality piece. And so we have done some research on that before we started having AI look at 360s. And there are things that you need to do to switch on confidentiality, especially if you're using sensitive inputs.

And then lastly, I just want to highlight this idea: the impact that I anticipate is the impact on you. And we already see this. How many of you use Google Maps a lot? How many of you have discovered that, in using Google Maps and all that, you’ve sort of forgotten how to get to places?

Right? So you either use it or you lose it—whether it's Google Maps or arithmetic or research abilities or whatever, etc.

And so the biggest impact is that AI can certainly amplify your output. It can also increase your skill of symbiotic partnering with this platform—or platforms. It can allow you to further develop your top-end skills, because you're able to reinvest, if you like, bottom-end work into top-end work.

And I guess this is the big thing that really has struck me over and over—and I’ve been thinking about a lot—is: Be careful what you outsource to AI. Sometimes there’s something to be said for carrying a heavy backpack up 7,200 steps.

And I don't know if anybody recognizes this picture on the screen. It’s from WALL-E. Yeah. I mean, I think it's a cautionary tale, right? And we can think of—in what ways have we already atrophied in some ways?

So I guess it's just being thoughtful and wise in itself about what we hang on to as human beings, as the AI continues to explode in terms of its capability.

Final Thoughts

And then, my last thoughts are—and again, I will work out a way if you would like to reference this, because I know there's a lot of stuff here—is:

What I found is what's really helpful to me is I think of AI as a super smart, literal, and naive intern who can't read my mind, but seems to have unlimited time and unlimited access to information, and who will provide me exactly what I asked for—whether I realize it or not—with a high probability.

So I offer that to you.

I think the next thing I would offer is: Ask AI for exactly what you want. Don't worry about how it gets there. Just ask for the result, in the way that you want it, for the use that you want it in. That's very helpful.

And then lastly, I would offer it as a three-step process:

1. Ask for it the best you can.
2. Review it.
3. Refine your ask—or even ask it to review some of your work. Ask it for feedback, etc.

Anyway, think of it not as a one-step but a three-step process. So instead of taking you 30 seconds, it might take you three minutes, potentially.

These are a couple of resources that I go to. I subscribe to The Rundown AI. It sends me something every day. It's very practical. It looks at the big picture and the small picture and all the things that are happening and all the platforms that are out there, etc.

If you want to go for a deeper dive, they have this “University” that you can click on and do a free trial of—which, as soon as I get half a chance, I’ll be signing up for.

And then lastly, there’s this Prompt Engineer GPT idea, which is: you don’t even have to be good at writing prompts anymore. AI will help you write better prompts to get you better, better impacts.