Video Transcript

Creating a Spark for Artificial Intelligence: Hear from IBM

So, now we're going to have you share your term definitions below in the comments, and feel free to
steal definitions that you see that you like or question definitions that you see and add to them to make
them more accurate and meaningful, add some examples. So, this is your opportunity to sort of
collaborate and crowd source some of those definitions. We believe that through this inquiry-based
learning process, it's really the experience of the term that gives it more meaning for you so that, when
you try to define it yourself, you have some ownership over that definition. This is also something that
you can do with your students. So, you can have them sort of experience any key terms that you'd like
them to learn, and then rather than memorizing a definition, they're writing their own definition. It also
gives you some insight into their level of understanding. So, if you're seeing definitions that aren't quite
what you're looking for, you can help them fine tune those definitions.

Now we're going to hear from our IBM content expert John Cohn, who is incredibly inspiring, and who
says I love making things, figuring things out, and working with my hands, but what I really love is
getting other people excited about making things, and you'll see this in action in his talk. So, I hope you
enjoy John Cohn.

Fascinating. I'm an AI, I would say, more or less an expert, and I'm learning quite a bit from this. So, I'm
really happy to be here, and I'm very, very honored to be with so many teachers. I just want to tell you a
little bit about myself is that I am an old and weird guy. I hope that's not a surprise to anybody, but I am
a total nerd. I grew up in the middle of the space program. I've been at IBM...I went to the nerdiest
school I could think of, which was MIT. I went to the nerdiest company, which was IBM, and I've
managed to be here for about 38 years. And just to give you a quick perspective, I want to introduce you
to a couple of other people. That is my late father-in-law, Gabriel Mariano. He worked at IBM for 36
years, and he loved making things. So, he, I had 38 years. He had 36. He liked making things like these
giant robotic machines, early, early, pre-AI, but they were highly automated machines for making big
computers. And he loved making things so much that he actually made my wife, or he helped make my
wife, who was also an IBM data analyst for about eight years. So, between us, we have about 82 years in
this company and 82 years in the industry.

And that gave me kind of an interesting perspective, and this is what I'm trying to share with you today is
that I kind of looked back quite a few years ago, a couple years ago, my company turned about a
hundred years old, and the industry that we were in turned about a hundred years old. I am a hardware
guy, and I love making things. I love making chips. I love making things that now think, and I looked
back over a hundred years to try to figure out what the pace of change and what was driving change and
try to see what insights I could get, not only on the technology, but on what it is to be a technologist or
what it is to be human in a time of great technology change. And what was really interesting is, if you
went back a hundred years, computation, you know, there were no computers per se, but there were
mechanical devices. As a matter of fact, our company first started out making mechanical scales that
would calculate how much you owed based on how much something weighed, but then we actually
needed to make that go faster, so we started making electromechanical. That was instead of somebody
turning the crank, a motor turned the crank, like the Mach-1 computer that was at MIT and Harvard. I'll
say MIT a lot this time because I love MIT. Go MIT. But that machine actually used a motor to be able to
calculate...to an amazing amount of calculations. It could multiply 50 10-digit numbers a second, which
was amazing. Then we went to vacuum tubes, which believe it or not, was what I was still studying a

1
little bit when I went to college back in MIT in the seventies. And when I joined in, there was something
called discrete bipolar transistors, and then CMOS circuits, and each one of these things, this part
actually shows how much compute you could get for the same amount of money, say a thousand dollars.
And this talks about how, what the pace of growth. You probably heard something like Moore's Law;
well, this is actually a hundred years of Moore's Law. And what it basically says is that the demand for
compute is so high and also our ability to increase it, and there's, you know, it's a great testament to
human innovation. But what is really interesting about it is that while it looks like a smooth curve, it's
anything but. It's all of these nice smooth lines of where things are working great and then a
discontinuity; and those discontinuities were something new started to work or something old stopped
working.

For example, back in the early 1980s, I was working on computers with bipolar transistors and they kept
getting faster and faster and they were the most amazing computers, but they started melting and
catching fire, and that was considered bad. So, luckily, we had another technology to move to, but you
see these gaps. Those gaps are really, really hard. When something changes, something new comes
along or something old stops working, it's the companies that are able to move through those and the
humans that are able to move through those disruptions are the ones who actually end up being
successful.

And if you look at about right now, one of the discontinuities here. So, most of the concentration on
compute for the last hundred years -- and that's a long time -- has been about how fast the computer is,
but I'll tell you what, all the stuff that we're hearing today about AI, that is by far the biggest change yet.
And I've been at this, like I say, myself for almost four decades, and my family for almost, for more than
eight decades, and this is by far the biggest thing that I've ever seen. It's big in the sense that it's making
things that were considered to be only something that humans could do, you know, possible, but it's
also generating a whole bunch of new opportunities, and I have to say, some new challenges, new things
that we have to be concerned about in terms of things like ethics and safety, et cetera. But if you think
about it from the standpoint of how much good this will do in the world, the current estimates are that
this will actually increase the GDP of the world about 16 percent, which is about $13 trillion, by 2030,
which is a huge good. Whatever your politics are, creating that much new opportunity, new value is
really, really good for the world, for the most part, if it's used responsibly.

Now, the interesting thing about that, though, is you as teachers, you are actually at the pointy part of
that, which is the students that you're teaching now are going to be entering the work force in a time of
unbelievable change, and even...and their parents are going to be in a situation where work is changing
quickly, you know, the nature of work. It is true that raising everybody's value, raising $13 trillion of new
economic value in the world is good, but just like those circles about having to make a disruptive change,
we're going to have to learn how to disrupt ourselves. So, the best thing I can tell you about how to
teach students, because the technology that we're seeing today is not the technology we're going to see
even in two years or six months. It's moving that fast. But what we have to do is we have to figure out a
way of teaching students curiosity, responsibility, ethics, hands on, and that will make them more
adaptive, more resilient and more passionate about this technology because, believe me, there is no
better time to have been a nerd. Your students are going to be entering a technology innovation wave
that's unlike anything that any of us have ever seen. It's just been, it's going to be fascinating, but we
have to teach them to be curious, we have to teach them to be ethical, we have to teach them to be
responsible, we have to teach them to be flexible. Not only that, but when they go home and talk to their
parents, it can help their parents get over this notion that they're going to have to adapt, too, because

2
the nature of work is changing, like manufacturing jobs are decreasing while service jobs go up. The
intellectual content of our jobs will actually go up even as we have computers to do it. But everybody's
going to have to change.

Now, what is the secret that I have in terms of how do you make students more curious, more
adaptable, more resilient? The answer is, maybe unexpectedly, is hands-on play. I mean, I've spent my
entire life...while I've had a really great career making chips for some of the fastest computers,
everything from super computers to video games. The kinds of things that have actually got me into the
current job I have now at IBM for MIT was because I was making weird things, like a 21 foot tall robotic,
wirelessly controlled robots, as part of Vermont's Own MIT club, which is called VOMIT, Vermont's Own
MIT club. This was a primitive way. It involved some very limited stuff. I wouldn't quite call it AI, but it
was my first hands on. This was something I did with a group of students. Here's another thing that I did
with a group of students. It's also 21 feet tall. It's a giant robotic remote controlled Ferris wheel for
carrying hippies across the desert at a place called Burning Man.

Now, why am I introducing that to you? Because this topic of making stuff like this is how I got to learn
and how I got passionate about it, that also allowed me to have great conversations with students of
every age, from kindergartners to 90 year olds. But what's interesting about it is that teaching this
hands-on play, when my company wanted to get into a new line of business, which is called Internet of
Things, I didn't have a portfolio. It was a new field that didn't exist and all I had as a portfolio was the
things I'd played for, 21 foot robots and 21 foot basically carousels and stuff that went on. I showed
them that, and they said, okay, you can run this business for IBM. So, I ran an almost billion dollar
business out of this beautiful building in Munich, and it was simply my portfolio and the skills that I
brought to it were done because I had actually used my hands, played, made mistakes, and figured the
stuff out because the technology that I was actually leading was very different than the technology that I
learned back at school, but what school gave me was this passion to go work with my hands. Not only
that, but when I was working in Munich just two years ago, that's when I fell in love with artificial
intelligence because trying to figure out how do you talk to things and figure out when they're going to
break, how to optimize them, how to use them safely, are they being used right? Well, the natural thing
is to use artificial intelligence, and just two years ago, that was kind of a novel idea of making a building a
giant artificial intelligence.

Well, as much as I loved Munich, commuting between there and my home in northern Vermont was a
little bit far, so I was actually able to use the same play, the fact that I had been using AI, to conjure up a
new job back at MIT, where it's actually like back to the future for me. It was because I was
experimenting, doing with my hands, that the new technology came along, and instead of being sort of a
relic, as the technology passed me by, I'm not able to lead, and I'm not saying I'm so great. The main
thing that I think I bring to this is great passion and just trying it out. If there's anything you can give your
students, it's that ability to go explore. That exploration is what got me. I am now, I guess I would say,
Yoda for the MIT IBM Watson AI lab at MIT, where we're looking at the next generation of AI algorithms,
where AI starts to get kind of closer to general intelligence, where it can actually reason and take on
tasks, broader tasks at super human speeds. New physics for AI so we can do it and be more
responsible, so we don't use up all the energy in the world trying to do AI. New applications for AI, and
specifically a lot of work on AI ethics to make sure that we're teaching people and teaching AI to make
good choices.

3
Now, as a teacher, you have a great opportunity to inculcate your students...is that a real word? You
know, to teach your students how to go after this on their own because you're not going to be able to
teach them the technologies of the future because it's changing so fast, but there are so many resources
out there. I don't know how many of you know about FIRST Rrobotics.

FIRST Robotics is a program that starts actually in kindergarten...what do you call that? K through five
and then goes on through middle school and then goes on through high school. It's designed as sort of
like a sports competition, but it's so much about technology. There's increasing AI component. It's very
hands on. It's very artistic. It's not just the tech kids; it's the theater kids and the drama kids and coming
together to actually learn how to work with your hands.

There's great groups that do code, like Girls Who Code, that actually have great resources for teaching
students as young as middle school how to do basic AI, like image recognition.

Or, Khan Academy, for those students who really want to learn about the new algorithms, you know,
before their professors do. There's so many resources out there. You don't have to be a genius. You
don't have to be an AI expert. You just have to be enthusiastic and pass on that notion that it's, you
know, that curiosity, responsibility, and play. I mean, I really think, I spend a lot of time trying to teach
our technology how to play.

[ TOY SOUNDS ] This is a tool that I just created with a bunch of students called Veremin. It's a visual
theremin. And if you...let's see if I can get it to play here. You basically can run it on a phone, and as
you're moving around with it, it will actually play somewhat beautiful music. Your students can actually
have the code. If you go to ibm.biz/veremin or go to that QR code, there is a tutorial down there that
shows how the code is built. And somebody who's like a medium advanced high school student would
be able to take that code, and she or he could actually go work with it. Does it solve the world's
problems? No, but it's so fun and it's so easy, and the techniques you learn on this are like many of the
techniques you've heard today.

So, the main message I would leave you with is make sure that you get that, yes, there's many things to
be concerned about with any new technology, but this is going to be really much more powerful and
good than it is bad. And the main thing you can teach people is the desire, the curiosity to go play with it
with their hands, to make a few mistakes, to learn from friends, to borrow, steal, kind of do it on their
own. It's just that. Teaching them how to play is the thing that is the greatest gift you can give them as
the technology changes out from under us at an incredible new pace. So, I'm super honored to be here
with teachers. I hope you really enjoy the rest of this course, and I'm looking forward to listening along
with you the whole time.

So, I love what John Cohn says about disrupting ourselves and learning alongside students. It's so
important, because the technology is changing at such a fast pace there's no way to get out ahead of it.
But the best thing is just to start your journey and start your students' journey and go on it together.