Inteligencia no humana

Translator: Gisela Giardino
Reviewer: Sebastian Betti

I’ve always been the nerd in the family.

Long before being a nerd was fashionable.

In the 90s my dad gave me
my first computer, a Commodore 64.

I spent so many weekends
copying video games into cassettes.

Because back then, programs and games

were copied to cassette tapes.

Nowadays, in times when everything
seems to be in the cloud,

seemingly intangible,
physical devices seem magical.

I thought they were magical at the time.
And they fascinated me.

When I grew up, that fascination
blended in with other passions,

like reading and studying.

When choosing which career
or trade to pursue,

I wondered what would allow me to do
what I liked most for the rest of my life,

which was to study.

That’s why I chose Philosophy.

Today, I’m a Doctor of Philosophy.

Two times I talked to my parents
about my future.

In one, I told them
I wanted to be a philosopher.

And in the other, I told them I was gay.

The two agree that they were
much more concerned about the first,

because they thought that
I was never going to get a job,

and that they were going
to help me all your life.

Luckily, I work since I was 23
and I’m very happy as a philosopher.

For more than 10 years I dedicated myself
to the philosophy of the mind,

which is the area of reflection
on how we think.

However, I never ceased to be
the son, the grandson, the nephew,

they resorted to when the computer
was running slow.

Neither, the boy who looked with interest

the specifications
of the latest smartphone,

with more or less the same passion
some of my friends

looked at Boca’s latest lineup
or the motor size of their last bike.

When those themes started to bore me,

I started wondering
about other aspects of the mind.

And a question popped up,
which many have done before.

Can machines think?

Siri: Could you say it again?
I didn’t hear you right.

That’s why today I study
the philosophy of technology.

Yes, because philosophers

we don’t just read Plato and Aristotle,

we also ask ourselves questions
about everyday life.

Perhaps the question of
whether machines can think

isn’t that amazing or that novel to you.

Well, that’s because nowadays
we hear a lot of people talk

about artificial intelligence.

We hear it on portals, on the radio,

even in the advertising of a bank.

Talking about artificial intelligence
is fashionable.

But it’s a very recent question.

The first one to formulate it
in a concrete way

was an Englishman, Alan Turing, in 1950.

He wondered if machines could think.

And the first thing he discovered was that
it depends on what exactly means to think.

And that is a topic not easy to agree on.

Today, when we talk about
artificial intelligence,

we usually refer to
machine learning algorithms.

They are those that allow us
to store and process

very large volumes of information.

Those algorithms do tasks
that we consider smart,

like recognizing patterns
or make good predictions.

They’re the ones behind,
for example, Alexa or Siri,

who can recognize our voice
and what we want to tell them.

They’re also behind those
incredible recommendations,

such as a Danish series
you’ve never heard of,

and we bet you will like.

Siri: Tomas, Siri has something special
for you today.

When you work with algorithms,

the important thing is not
what we want them to do,

but what you ask them to do.

Let me share a real story.

A group of scientists fed an algorithm

with all the information on a flight:

type of aircraft,
number of passengers, route details.

And they asked it to tell them
the most efficient way to make it land.

The algorithm gave its verdict.

The most efficient way was to collapse
all systems in mid-flight

and let it crash.

It’s logic was impeccable.

An algorithm, a computer,
doesn’t do what you want them to do.

It does what you ask them to do.

Let’s do an experiment.

One of the most successful
areas of application

for artificial intelligence algorithms

is city navigation.

I’ve been living in the city for 20 years

and still today I get lost at some streets
or in certain neighborhoods.

That’s why I asked
a group of programmers

to develop a system,
absolutely customized for me,

so I don’t get lost anymore.

Let’s try it out.

What is the best route to get
from Plaza de Mayo

to the Teatro Colon?

Siri: The best route to get from there
to the Teatro Colon

is a straight line from
this point to that building.

The logic is impeccable.

However, I’m going to hit
the first building

I want to go through
walking in a straight line.

Let’s try again.

What’s the best route to get

from Plaza de Mayo to the Teatro Colon?

Siri: The fastest route from here
to the Teatro Colon

right now is to take Diagonal Norte,

turn right on Esmeralda,

and then at the crossing with Lavalle
take Corrientes Avenue

and walk until that building.

Well, that’s a good answer.

It happens, for reasons beyond this talk,
that I don’t like Lavalle very much.

So let’s see if it gives me
a better answer.

What’s the best route
to get from Plaza de Mayo

to the Teatro Colon?

Siri: Walk straight on 25 de Mayo
until Lavalle,

so you don’t cross that records store,
you used to go to with your ex,

that you still pass today
and makes you sad.

OK, lots of personal information.

However, it is a good answer.

We can confirm then that defining
what intelligence is not easy.

And that there’s no single answer.

To determine if something
or someone is smart

depends on the subject we are analyzing

as much as the one
making the evaluation.

This brings me to a question
spinning in my head

for a long time.

Can we create an intelligence
different from ours?

There is no doubt that
machine learning algorithms

are changing our lives
in a profound way never seen before.

But I’m not sure we made any progress
in answering Turing’s question.

We have created computers and systems
that are incredibly fast

for a lot of tasks.

But they don’t solve them
necessarily better.

They just solve them faster.

Current debates in philosophy
on artificial intelligence

revolve around how our mistakes

are repeated and perpetuated
by algorithms.

Today’s devices are as wise
and as dumb as we are.

However, they are
more powerful and faster.

This brings me to another question.

Can we recognize an intelligence
different from ours?

Humanity doesn’t have an honorable history
of dealing with the different.

Almost a hundred years passed

since Columbus stepped on our continent

and Europe determined that
the inhabitants of America were people.

Today, the movements against
animal cruelty are teaching us

that there can be non-human people.

History never ceases to amaze us.

And we have to start discussing
the existence of intelligences

that may be different from ours.

Since we don’t have a good record
of treating the different,

my invitation today is for us to start
thinking about other intelligences.

If we are actually going
to have difficulties

in recognizing intelligences
different to ours,

what if they already exist
but we haven’t recognized them yet?

Look around you.

What if we’re surrounded by human
and non-human people and we don’t know it?