0:00:18.000,0:00:23.560
In 2016, I was awarded a prize by the[br]National Forum for Teaching & Learning

0:00:24.320,0:00:27.700
supported and sponsored[br]by the Ministry of Education in Ireland,

0:00:27.700,0:00:30.970
peculiarly named:[br]"National Teaching Hero."

0:00:32.180,0:00:33.680
The reason for this award

0:00:33.690,0:00:36.820
was my availability towards my students

0:00:37.020,0:00:38.250
and my capability

0:00:38.250,0:00:42.640
to create less formal and more [br]comfortable educational environments.

0:00:43.160,0:00:46.240
Let's imagine this place as[br]a large university classroom.

0:00:47.880,0:00:49.990
I'm used to enter the classroom,

0:00:49.990,0:00:54.560
[in the] first ten minutes,[br]when students enter and take place,

0:00:54.560,0:00:58.960
I plug my computer to the speakers[br]and turn a classical music on.

0:01:00.200,0:01:03.080
I think this is the first step

0:01:03.080,0:01:07.480
to build less formal and [br]more comfortable educational environments

0:01:07.480,0:01:10.440
and try to keep the attention[br]of students at a high level.

0:01:12.371,0:01:14.340
Unfortunately, this is not always easy.

0:01:16.560,0:01:22.640
In my lessons, I employ a method in use [br]since the ancient times: storytelling!

0:01:24.250,0:01:25.540
Pedagogically speaking,

0:01:25.540,0:01:30.520
storytelling is a method[br]based upon the use of narratives,

0:01:30.520,0:01:33.200
aimed at transmitting [br]knowledge to students.

0:01:34.040,0:01:37.800
I would start my lesson[br]exactly with this method,

0:01:38.040,0:01:43.120
by explaining, describing a topic[br]on everyone's lips nowadays:

0:01:44.274,0:01:45.770
Artificial Intelligence.

0:01:47.786,0:01:50.220
Like every story narrated to children,

0:01:50.280,0:01:54.150
I'd like to begin my story with [br]"Once upon a time."

0:01:56.760,0:02:02.210
Second World War, 1942, United Kingdom,

0:02:03.160,0:02:04.500
Bletchley Park:

0:02:05.130,0:02:09.030
a mansion house in the north of London.

0:02:10.479,0:02:13.920
There was a thirty years old guy: [br]Alan Turing.

0:02:15.680,0:02:18.040
Alan graduated at King's College,[br]Cambridge,

0:02:18.670,0:02:20.860
and he obtained [br]a research doctorate in Logic

0:02:20.870,0:02:23.770
at Princeton University, in the USA.

0:02:24.880,0:02:30.120
At that time, the Germans made use[br]of a special machine: Enigma.

0:02:32.040,0:02:34.470
It was like a typewriter:

0:02:34.470,0:02:36.370
the operator typed some keys,

0:02:36.370,0:02:41.090
but, instead of printing those letters[br]on a paper sheet,

0:02:41.090,0:02:42.960
other letters were printed,

0:02:42.960,0:02:46.600
according to an encoding mechanically[br]set under the machine.

0:02:47.440,0:02:51.760
The Germans used this machine[br]to communicate with each other.

0:02:52.920,0:02:55.550
Anybody had listened in on this sheet

0:02:55.550,0:02:57.930
had in front a meaningless text.

0:02:58.950,0:03:00.030
It was encrypted!

0:03:02.160,0:03:07.880
Alan Turing was one of the leading figures[br]at Bletchley Park:

0:03:08.880,0:03:12.810
he and his team implemented a machine,

0:03:12.810,0:03:15.320
the one you can see behind me,

0:03:15.320,0:03:20.320
able to decipher the texts [br]written by the German.

0:03:21.680,0:03:23.310
Due to this invention we believe

0:03:23.310,0:03:27.480
that the war terminated two years earlier,

0:03:27.480,0:03:28.820
saving many human lives.

0:03:30.480,0:03:34.160
After the war Alan Turing continued[br]his research in Logic

0:03:35.840,0:03:39.200
and he is considered[br]the father of Computer Science,

0:03:39.200,0:03:41.900
the father of Artificial Intelligence.

0:03:42.720,0:03:47.290
With his Turing machine,[br]he formalises the concept of computer

0:03:47.290,0:03:50.500
even before the computer[br]was actually built.

0:03:52.480,0:03:57.280
In 1950 he published a paper[br]on the journal "Mind":

0:03:57.280,0:03:59.290
"Computing Machinery and Intelligence,"

0:04:00.220,0:04:03.400
where he proposed the Turing test.

0:04:03.400,0:04:07.610
The question behind the Turing test[br]is a well-defined one:

0:04:09.170,0:04:11.380
Can machines think?

0:04:13.560,0:04:16.110
It is at that time that[br]Artificial Intelligence begun.

0:04:18.730,0:04:21.920
Probably most of you have watched [br]the movie "The Imitation Game":

0:04:21.920,0:04:23.770
The game of imitation.

0:04:23.770,0:04:25.530
I am going to describe it to you.

0:04:26.840,0:04:31.480
Let us suppose that a person is here,[br]pressing keys on a computer keyboard,

0:04:32.400,0:04:34.540
who asks, pose some questions

0:04:34.540,0:04:36.750
and on the opposite side of the computer

0:04:36.760,0:04:39.730
there is a machine M and an operator O.

0:04:41.670,0:04:43.560
Alternately,

0:04:43.560,0:04:48.520
the machine M and the operator O[br]answer to the person P.

0:04:50.360,0:04:54.840
It is said that the machine M[br]passes the Turing test

0:04:55.480,0:04:58.290
if the person P is not able to understand

0:04:58.290,0:05:00.720
when the answers come from the machine

0:05:00.720,0:05:03.780
and when the answers [br]come from the operator.

0:05:03.801,0:05:05.670
This is The Imitation Game.

0:05:07.200,0:05:10.923
The machine must have special features

0:05:11.203,0:05:12.980
in order to pass the Turing test.

0:05:14.160,0:05:16.990
It has to interpret natural language:

0:05:16.990,0:05:19.050
the question asked by the person.

0:05:19.800,0:05:22.960
It has to represent knowledge[br]in order to formulate answers.

0:05:23.840,0:05:27.440
It has to think in an automatic mode[br]in order to formulate such answers.

0:05:28.480,0:05:31.050
And it has to learn automatically.

0:05:33.120,0:05:35.190
There are many approaches

0:05:35.190,0:05:37.200
to study Artificial Intelligence.

0:05:37.640,0:05:43.240
One of them is the cognitive approach:[br]it is based upon the human thinking.

0:05:45.280,0:05:46.730
According to this approach,

0:05:46.730,0:05:49.480
there are two ways to study[br]the human thinking:

0:05:49.480,0:05:53.520
either we try to capture thoughts[br]right when they occur,

0:05:54.095,0:05:57.000
or we try to model thoughts[br]at a psychological level.

0:05:59.154,0:06:01.530
For this reason, [br]we say that Artificial Intelligence

0:06:01.530,0:06:05.840
is closely connected to neuronal and[br]cognitive sciences and to psychology.

0:06:06.110,0:06:08.630
According to this approach,[br]the assumption is that,

0:06:08.630,0:06:10.240
if we can have

0:06:10.240,0:06:13.150
a true representation[br]of the human thought,

0:06:13.150,0:06:15.500
then we can transfer it to a machine.

0:06:17.760,0:06:21.260
Another approach is the one[br]based upon the laws of rational thought.

0:06:22.520,0:06:27.240
Probably most of you has heard about [br]the Aristotle's syllogism.

0:06:28.980,0:06:30.570
Socrates is a man;

0:06:31.740,0:06:33.510
all men are mortal;

0:06:34.550,0:06:36.380
[then] Socrates is mortal.

0:06:36.920,0:06:39.810
This is a deductive reasoning:

0:06:40.660,0:06:43.410
If we have two truthful premises,

0:06:43.760,0:06:45.810
we can infer a truthful conclusion.

0:06:45.900,0:06:48.340
Deductive logics comes from here.

0:06:48.944,0:06:51.530
According to the laws [br]of rational thought approach,

0:06:51.880,0:06:56.960
we try to build deductive arguments[br]and to transfer them to a machine.

0:06:59.600,0:07:03.470
Another approach is [br]the rational agent approach.

0:07:04.840,0:07:07.800
A rational agent, an entity,

0:07:08.510,0:07:10.510
has to act,

0:07:10.800,0:07:13.480
has to adapt itself to the context,

0:07:14.110,0:07:17.990
has to fix goals to itself,[br]and be able to carry them out,

0:07:18.480,0:07:21.240
and it has to act in a rational way.

0:07:22.480,0:07:26.400
Therefore, the Turing test[br]is related to intelligent agents.

0:07:27.040,0:07:30.360
By rephrasing the Turing statement,[br]"Can a machine think?",

0:07:31.080,0:07:35.280
we can now say: is it possible to build[br]a machine, an artificial agent,

0:07:36.320,0:07:38.180
able to think,

0:07:38.180,0:07:41.660
able to show understanding and rationality?

0:07:43.600,0:07:45.780
Artificial Intelligence, therefore,

0:07:45.780,0:07:50.200
aims at developing artificial [br]intelligent entities.

0:07:50.200,0:07:52.510
Your mobile phone is an entity.

0:07:54.720,0:07:57.600
By developing artificial entities

0:07:57.720,0:08:02.090
we try to understand intelligence[br]as a psychological construct.

0:08:02.504,0:08:04.750
Once we have a better knowledge[br]

0:08:05.260,0:08:08.080
about this concept,

0:08:08.320,0:08:11.360
we try to develop[br]artificial intelligent entities

0:08:11.360,0:08:14.040
to support humans: it's a cycle.

0:08:15.980,0:08:20.980
But let's see now whether[br]machines are able to think.

0:08:21.480,0:08:24.570
I'd like to briefly describe you

0:08:24.570,0:08:27.570
the state of the art [br]of Artificial Intelligence

0:08:27.570,0:08:30.319
and I'd like to use five classes

0:08:30.319,0:08:34.790
to classify artificial agents[br]according to their abilities:

0:08:35.808,0:08:39.520
we have sub-human and par-human agents,

0:08:39.520,0:08:43.799
over-human, super-human agents[br]and then we have optimal ones.

0:08:44.068,0:08:46.470
I want to explain them [br]with some examples.

0:08:46.800,0:08:51.040
Optimal agents are the ones[br]which act better than all the people

0:08:51.349,0:08:53.070
and you can't do better than that.

0:08:53.280,0:08:55.800
For instance, agents[br]which solve the Rubik cube,

0:08:55.800,0:08:59.260
those that play at "Four in a row"[br]in the best way,

0:08:59.260,0:09:00.400
or at Tic-tac-toe.

0:09:00.400,0:09:06.000
Consider that some years ago, a boy,[br]given an initial state of the Rubik cube,

0:09:06.000,0:09:08.800
solved it in 4.73 seconds.

0:09:09.373,0:09:11.770
Some months ago an agent, [br]a robot was developed,

0:09:11.770,0:09:14.990
which can solve it in 0.63 seconds.

0:09:16.840,0:09:22.160
We have super-human agents [br]which act better than all the humans,

0:09:22.740,0:09:26.510
for instance in the chess game,[br]in the Scrabble game.

0:09:27.170,0:09:30.730
Some years ago the chess Russian [br]champion Kasparov

0:09:30.730,0:09:33.500
was defeated by an artificial agent.

0:09:34.480,0:09:36.610
We have over-human agents,

0:09:36.610,0:09:40.160
the ones which almost act better [br]than most of the humans,

0:09:40.160,0:09:45.160
for instance in the Texas hold 'em poker,[br]in answering the Quiz Show questions.

0:09:46.280,0:09:48.343
We have par-human agents,

0:09:50.330,0:09:54.500
which act almost like all the humans,

0:09:55.113,0:09:57.578
for instance in cognitive activities,

0:09:57.578,0:10:01.260
such as crosswords or image classification.

0:10:01.730,0:10:06.200
Finally, we have sub-human agents,[br]which act worse than all the humans.

0:10:06.200,0:10:12.250
Examples include objects classification,[br]handwriting recognition,

0:10:12.520,0:10:16.120
vocal recognition, translation[br]from a language into another one.

0:10:17.770,0:10:21.470
But if there is something that artificial[br]agents nowadays are not able to do

0:10:21.470,0:10:23.920
is for instance, disambiguation:

0:10:23.920,0:10:28.760
Are we talking about the apple as a fruit,

0:10:28.760,0:10:32.480
or are we talking about the brand[br]of the Apple Corporation?

0:10:33.440,0:10:35.850
And one thing that agents[br]are not able to do

0:10:35.850,0:10:39.560
is reasoning in the real world[br]under situations of uncertainty.

0:10:40.520,0:10:44.910
These are the main limitations[br]of Artificial Intelligence,

0:10:44.910,0:10:48.780
and because of these, it is believed that[br]we are far away to pass the Turing test.

0:10:49.600,0:10:55.570
Now, let's try to understand [whether][br]machines will be able to think.

0:10:57.560,0:11:00.880
Some years ago, in America,[br]a concept has been coined,

0:11:01.160,0:11:03.280
named "Technological Singularity":

0:11:03.430,0:11:08.230
by Ray Kurzweill, a world-renowned [br]expert in Artificial Intelligence.

0:11:09.091,0:11:10.930
Let's imagine a timeline.

0:11:12.480,0:11:17.270
Let's imagine a line indicating[br]the human intelligence, increasing.

0:11:18.320,0:11:23.050
Let's now imagine a red line[br]indicating the machine intelligence,

0:11:23.050,0:11:25.140
with an exponential trend.

0:11:25.800,0:11:30.880
This trend follows the Moore's law,

0:11:31.440,0:11:34.980
whereby the computational complexity,[br]for instance,

0:11:34.980,0:11:38.660
as measured by the number of transistors [br]embedded in a chip,

0:11:38.900,0:11:42.810
doubles every two years[br]and quadruples every [three] years.

0:11:43.720,0:11:47.700
According to Ray Kurzweill, [br]in 2010 we should have been able

0:11:47.700,0:11:52.194
to use this computational complexity

0:11:52.194,0:11:54.174
to emulate the human brain --

0:11:54.174,0:11:55.780
I didn't see anything.

0:11:56.480,0:11:59.840
In 2020, with 1,000 dollars[br]we will have access

0:11:59.840,0:12:02.730
to this computational capacity.

0:12:03.960,0:12:07.020
In 2025, according to Ray Kurzweill,

0:12:07.020,0:12:11.110
we will be able to scan our brain [br]in a very accurate way.

0:12:11.960,0:12:14.220
And eventually, in 2029

0:12:15.120,0:12:18.120
machines will pass the Turing test!

0:12:19.930,0:12:24.400
And then, in 2045,[br]he refers to that point in time

0:12:24.400,0:12:27.510
when the technological singularity[br]will happen,

0:12:29.240,0:12:32.720
when the machines, machine intelligence,

0:12:32.720,0:12:34.830
will follow an exponential trend

0:12:34.830,0:12:38.720
that will significantly affect[br]the human intelligence.

0:12:40.760,0:12:45.680
In his paper, published [br]on the journal "Mind,"

0:12:45.680,0:12:48.680
Turing not only proposed his test,

0:12:48.680,0:12:52.780
but he also suggested nine objections[br]against his own test.

0:12:53.400,0:12:58.100
These objections are nine objections[br]against Artificial Intelligence.

0:13:00.400,0:13:04.360
Some years ago, when I was a student[br]at the University of Varese,

0:13:05.200,0:13:09.320
I attended a course on "Epistemology,[br]Deontology and Ethics in Computer Science"

0:13:09.320,0:13:12.960
held by Prof. Gaetano Aurelio Lanzarone.

0:13:12.960,0:13:15.960
Unfortunately, he passed away [br]some years ago.

0:13:15.960,0:13:22.300
One of the assignments we had to do[br]was to propose a tenth objection

0:13:22.560,0:13:26.000
against the Turing test,[br]against Artificial Intelligence.

0:13:27.120,0:13:30.300
I was the only one who proposed [br]a tenth objection

0:13:30.520,0:13:32.900
expressed as a mathematical equation,

0:13:34.560,0:13:36.640
that I labelled "human stupidity."

0:13:37.490,0:13:39.980
I'd like to explain it in simple terms.

0:13:40.000,0:13:44.300
Let's assume we take[br]the intelligence of all humans

0:13:44.300,0:13:45.900
and we put it all together.

0:13:46.070,0:13:48.610
The Sum symbol of the equation[br]on the left hand side.

0:13:49.729,0:13:52.920
And we transfer this intelligence [br]as a whole to a machine.

0:13:54.000,0:13:57.610
Then, we get an equality of intelligence.

0:13:57.610,0:14:00.860
But in some way the machine[br]becomes more intelligent than us.

0:14:01.910,0:14:04.110
Though, if it is true that were us

0:14:04.600,0:14:08.653
who have transferred our intelligence[br]to a machine

0:14:09.313,0:14:12.313
and it becomes more intelligent than us,

0:14:12.313,0:14:13.333
it is also true, as well,

0:14:13.333,0:14:16.900
that we let the machine become [br]more intelligent than us.

0:14:19.160,0:14:21.500
Then, in order to conclude my story,

0:14:23.160,0:14:28.430
and referring back to the initial [br]question: "Can machines think?",

0:14:29.640,0:14:33.516
I'd like to leave you with[br]an open question:

0:14:36.321,0:14:39.511
Does really makes sense [br]for us to let them think?

0:14:40.751,0:14:41.741
Thank you.

0:14:41.841,0:14:43.911
(Applause)