字幕表 動画を再生する
TRACE DOMINGUEZ: Hey, everyone.
Thanks for tuning in to Seeker Plus.
I know this is also on Seeker.
This is a slightly different format
where we get really deep into a topic for a little while.
So stick with us.
I'm Trace.
This is going to be really cool.
We're talking about artificial intelligence today.
And we didn't want to present it as artificial intelligence is
scary or artificial intelligence is the best,
but come down in the middle and give you both perspectives
because there are people on both sides of pretty much every one
of these debates.
So let's kick into it.
Artificial intelligence is everywhere,
and it's growing both in scope and also in scale.
It's just getting to be in everything.
And people have strong opinions on this.
One article called "The Debate Itself--
Singulatarians Versus Skeptics," which
we thought was kind of cool.
And people do see artificial intelligence
as a possible threat--
people like Elon Musk, Stephen Hawking,
Bill Gates kind of, and at one point Steve Wozniak.
On the other side, people see artificial intelligence
as the future and inevitable-- people
like Larry Page of Google-- and said that in order for Google
to achieve its mission, that means achieving AI.
Mark Zuckerberg uses AI to run his home,
and Facebook is moving towards having AI assistants.
Even Uber has an AI division.
So it is like this old tech new tech and with Elon Musk
right in the middle.
But chances are you all probably know this stuff already
because you follow Science News likely
if you're watching this show.
But one way to think about this is positives and negatives.
Everyone has an opinion about AI, and both of those opinions
are usually valid.
So we thought we'd see how the debate is put together
rather than coming down on one side or the other.
So let's quickly define artificial intelligence.
It's not just robots and chat bots and assistants.
It's being used across all sorts of fields in industries,
and it's computers and it's machines
imitating human intelligence.
But it's not just that.
It's more than just learning.
It's more than just replicating our intelligence.
It's about learning new things, and it's also
about the machines learning on their own.
The big idea, the big concept here could be great,
but it also could be kind of scary.
So let's talk about different ways
that you could apply artificial intelligence.
A big one being talked about right
now is artificially intelligent automobiles, trucks,
and cars that drive themselves, which, if you think about it,
that is an intelligent task.
It requires decision making and learning and learning
from your experiences.
And a big thing would be safety.
90 percent of car accidents now, car crashes,
are caused by humans, human error.
Smart cars, intelligent cars, they
would be able to take in the environment around them.
GPS could tell them what roads they're on and also
what buildings are nearby and other monuments and things,
of course.
They have cameras and scanners.
So they can see the trees and the way the road actually moves
versus what it's supposed to do.
And you can see things like other cars.
Which what if they were smart, too?
Then they could interact with each other.
And then you have this network that is moving everywhere.
What if you have smart traffic lights?
Eventually this whole system could
be one giant AI system where they're all
talking to each other.
And then we could virtually eliminate 90% of crashes,
right?
Assuming we're eliminating all of the human error ones.
Of course, there are moral issues here.
The car might be designed to cause the least
damage to the owner of the car.
It might not want the car to be destroyed in a crash.
And so that's a moral question.
Can the owner of the car, say, a taxi company
decide not to protect passengers or predestines
over protecting their property?
This is an ethical dilemma.
What about a guy who owns a car?
He would want to protect his family at all costs.
But what if in doing so it sacrificed
other people's property or destroyed
other intelligent cars?
It gets complicated because nothing
is black and white when you get it out into the real world.
And that's part of the problem with artificial intelligence
across the board right now--
is that the real world is messy.
Is it worth it to have a self-driving intelligent car
that might not think some humans are worth saving
over other humans or that some property isn't worth saving
over other property even if it can prevent more accidents,
even if it knows the fastest routes and could drive itself,
so you wouldn't need to worry about people under age
or people over age or people who were
too intoxicated or inebriated in any number of different ways?
At the end of the day, intelligent automobiles
mean that, on the good side, we don't have to go to the DMV.
We don't have to worry about drunk driving.
But on the bad side, we're taking our control away
from something that happens all the time all around us
especially in urban areas.
So if you're in a car and it's raining and you slide,
you go around a corner and maybe something bad happens.
The road's too wet.
The car did something wrong.
It's either going to destroy the life
of the person in the car or a pedestrian on the sidewalk.
That decision is made by a machine not by a human
and maybe by somebody who programmed
or owned that machine.
We as a society have to make a decision
to give up that control and hand it to an artificially
intelligent machine.
How do you feel about that?
It's a debate.
There's good and bad on both sides.
And this applies across the board.
Let's go to another example--
marketing.
Marketing is pervasive, especially
in the era of the internet.
Marketers want to know where you're going to be,
what you're going to buy, and what it takes
to get you to buy something.
You shop for shoes that one time,
and now there's an ad for shoes on every website you
visit and that shoe and a couple of other different shoes
in a variety of true colors.
And they're all really nice, and you kind of want all of them.
But you kind of want none of them
because they're just everywhere.
Or maybe you regularly visit a site on your computer
before you buy something.
Maybe you visit it 20 times.
Now right now all of that stuff happens and all of that data
is there, but no one's really looking
at all of that data except artificially
intelligent machines.
Predictive shopping and also recommendation engines
on Netflix and on Amazon--
those are basic artificial intelligences.
So imagine if they got way better.
When you shop for something, do you just go and buy it
or do you visit the same site once or twice, then
maybe look at some reviews, then maybe ask your friends about it
and then go back and buy it?
An artificial intelligence system
might be able to get you to buy that sooner because they
know your behavior.
So when you go there the first time,
what if an artificial intelligent system
just pulled in recommendations from the sites that you already
visit that it knows you visit because of the cookies--
little files-- that all of those sites
have stored on your computer?
What if it knew your social network
and it went to your friends who it knew