When you're learning new material, like

Barb said, it's really important to have different problems, to work on

different questions that you can go and try to answer to understand the material
at a deeper level, to practice, to think through different important issues and
concepts and things that you need to know. But often those are, you know,
hidden behind textbooks and answer keys. And how do we get to those things? And
when you're learning a new concept
of or some idea or you're listening to a presentation for the first time, it can
be hard to formulate those questions. And so how do we go and fix that? With
generative AI, one of the things we can use is we
can use it to generate questions. And this is such a simple concept, but one that
gets missed when we're
running off to get answers. So let's take a look
at what it looks like. But I want you to think deeply about how
you get it to generate questions, and I don't want you to just go and ask it to
generate questions
in a prompt that this big. I want you to think about
giving it rich source material. So one of the things you can do in
most of the tools is you can go and attach files to the prompt. And so if you go
into ChatGPT, you'll
probably see something like a plus or a paperclip. If you go into Claude, you'll
probably see something very similar. And that's your ability to go and
upload and give it additional information
to go with the prompt. When you do this, your prompt becomes not
only that file, but also the question or instructions that you're giving the model.
And this is really, really helpful. So what I've done here is I've taken
one of my PowerPoint presentations. You can imagine this is the slides for
a training or a course that you're taking, and you want to go and
understand it at a deeper level. Or maybe it's a new proposal that you
received from an outside vendor who's going to do some services for you, but it's
in an area that
you don't know a lot about. And so the starting point of helping
ourselves to understand the material, to practice the material, is to be able to
generate
questions related to the material. So I've gone and uploaded my presentation,
it's about 100 PowerPoint slides. And I say,
help me understand this material better. What are five questions I should ask
about it to deepen my understanding? And this is my own presentation, so
I think it's quite fun to do this. And it comes back and it says, what are the key
advantages of using
prop patterns in generative AI? And how do they improve the quality of
outputs compared to unstructured prompts? Now, if you've heard me talking
about prompt patterns already, you know this is my presentation,
and that's a good question. And being able to answer that
question can be very helpful. How does the concept of augmented
intelligence differ from traditional artificial intelligence? And augmented
intelligence is a concept
that I talk about in this presentation. Not of using artificial
intelligence to replace people, but people working together with
artificial intelligence to augment and amplify their own critical thinking and
reasoning and creativity. Of ways of using it to enhance
what they do, but not replace it. And it goes on with a number of questions. And
all of these questions are questions
that if somebody can answer these questions about my presentation and
about the material covered there, it will deepen their understanding. They will
know it more if they
think about these questions and they think about the answers to them. It's very
helpful. And so this simple idea of generating
questions based on material that we're learning about, but
not just going in and saying, hallucinate a bunch of questions
that might possibly be related. But actually taking the materials that we
are working with, the actual slides for the class, the actual syllabus,
readings, proposals, presentations, and building questions from
them is the way to go. Because it helps reduce this hallucination
that we're talking about, but also helps us to really target
the information that it's analyzing in order to produce the questions. Let's go and
look at another example. It's not just questions about
the materials that we're learning in a class or a course. It's not just a written
proposal, but
often what we have is teams of people that are working through ideas together, that
are trying to solve some problem together, and they're learning in the process. And
so one of the things we can do is we
use all of these online meeting tools like Zoom and Microsoft Teams, and we can
record the meetings in
there and get transcripts out. So I've taken a transcript that I had from
a discussion with a close collaborator. It's a company that we're
collaborating with. I've actually done some courses
with the CEO of this company, and I took the transcript from
the Zoom call we had, and I said, without listing any names, what are five
questions our team should ask and discuss to help think through our plans? And it
comes back and it says, what
are the core objectives and priorities for this collaboration? How can we
effectively engage and
align all stakeholders involved? And it's giving me things that we
should learn about and think about. It goes on, and it says, what
are the most significant challenges or opportunities within our current
processes that innovation could address? How should we brand and
communicate what metrics, you know, or key performance indicators will
be used to measure success? All of these things are important
questions that we should think about as a team and answer, because we're
actually going on a journey to learn together about how to solve this problem,
how to work with them, to incorporate generative AI in
the space that they're working. And there's a lot of complexity there. So this is
actually bringing questions in
that the team as a whole can discuss to help learn together about how we're going
to approach this problem and solve it. So we don't have to just think about
questions for the materials, the written materials, but we can also think about
questions that teams as a whole should ask about how they're coordinating, planning
and communicating and working together. Or another example is we
just want more practice, we want to go and practice things. We've seen the
importance of practice, not
giving up and giving over everything, but we really need to be practicing. How do
we get more problems
to directly practice with? And so we can take something like the
slides from a course, or slides in this case from a talk that I'm giving and
say, what are five questions that, if I can answer them, would show that I have
a broad understanding of this material. And so when we create this, we're creating
a set of questions that I
can then practice with and think about. And go and begin answering to practice,
because it's not just about thinking. We can use questions to help us think
more deeply and learn by going and directing how we go and
think about a particular problem. We can use it in sort of questions to
help a group go and learn together and dissect a problem and think about what
might be important to talk about. But we can also go and use it for
actual direct practice and get quizzes, get problems that we can try out,
all kinds of things like that. And we see some great examples here. What are the
two key revolutions in
computing driven by generative AI? And how do they differ from previous
computational interfaces and capabilities? If you were in a class
listening to me lecture, that would be a fantastic
question to go into a quiz. But that's exactly the type of
thing that you can directly use and equip yourself with is take any
material and turn it into practice. Rather than taking material and turn
it into answers, take the material and use it for practice to generate as many
different questions as you want to base those questions in particular domains. You
can stay focused on these aspects
of the presentation, generate questions related to these topics, or test me and
tell me what topics I'm weak on. Give me a bunch of questions, tell me how
I did all kinds of fascinating things that we can use to enhance our own practice,
and generating questions or problems is one of the most
basic things that we can do.