Intelligenza Artificiale: manipolazione delle menti e futuro della scienza

Domande al Nutrizionista
25 iscritti
1 Visualizzazioni video·22 set 2023
L'intelligenza artificiale ha fatto negli ultimi anni passi da gigante, aprendo un abisso di dilemmi ma anche di incredibili potenzialità: quali potrebbero essere le ricadute di questa straordinaria tecnologia nell'ambito della scienza nutrizionale? Visita il nostro sito https://www.spaziosfera.com
Seguici su Instagram https://www.instagram.com/spaziosfera
Seguici su X https://www.twitter.com/spaziosfera
Seguici su Facebook https://www.facebook.com/spaziosfera
A greeting to you friends of the web, today's video is dedicated to a very current topic,
that of artificial intelligence applied to the theme of our channel, that is nutrition.
The question is essentially what could be the positive outcomes of this new technology,
as long as there are, in improving our diet?
Well, I am convinced that, as there could be many, and I see at least three very concrete ones that we will talk about today.
Before getting there, however, in the first part of the video, let's do a little introduction to the subject,
because even if there is a lot of talk about it lately, not everyone has very clear ideas about the potential of artificial intelligence.
When challenges related to artificial intelligence are debated, I very often hear,
yes, but in any case, the robots are us who program them, they do what we tell them.
Here, perhaps some have escaped the last 5-10 years of advancement of technology in this area.
Therefore, first of all, I summarize the example that is always made to explain what the difference is,
that is, the comparison between two events in the history of artificial intelligence.
One in 1997, when the IBM Deep Blue robot won against Garry Kasparov, the world chess champion,
and the other 20 years later, in 2017, when AlphaGo from Google won against Ke Jie, the world champion of Go.
They seem two completely comparable events, but in reality the difference between them is abysmal.
The robot that won by a landslide, in fact, was programmed by us, in the sense of us humans, the IBM programmers.
With all the possible moves to make in various situations, but it was a classic programming,
if then, if the opponent does this, what do you do?
Simply, he had a power, an incredibly superior calculation speed compared to the human brain,
so he could analyze one by one all the possible combinations for each move and predict all the possible results,
and therefore he often won.
And when the human chess champion studied the game, he said,
oh, this move is intelligent, why didn't I think of it?
Quite the opposite with the game at Go, 20 years later.
Go is that ancient Chinese game that you play on the grid of 19 lines by 19,
and it is infinitely more complex than chess, so much so that here the possible combinations are almost infinite,
they are 10 raised to the 360th, which exceeds the possible combinations of chess of 249 orders of magnitude.
So it would have been absolutely impossible to program the robot as it was done for chess,
it would not have been enough a billion programmers committed for the entire duration of their life.
And in fact, not only the Google robot was not programmed with all the possible moves,
but the rules of the game were never explained to him,
all he did was look for months at the champions who played Go and tell him at the end of each game who had won.
Using the new technology of deep learning based on neural networks,
that is, in the structure of the neural networks of the human brain,
the Google robot, simply seeing a lot of games,
not only inferred the rules of the game, but also elaborated the strategies and ways of playing to win.
Ok, so no one ever told him if he moves like that, you move like that, he did everything by himself.
Then he played with the champions and won and won and won again.
And this time, when the champions studied the game afterwards, they could not understand anything,
the robot's moves did not seem to make any sense to them.
Yet with those moves the robot won.
So in this case the robot went beyond human intelligence,
he understood things that we did not understand,
he saw things about that game that the best champions in the world were not able to see.
And so in this case artificial intelligence really surpassed us,
in the sense that it made the right decisions, but acting in a way that was completely incomprehensible to us.
Ok, so the potentials are at this level.
Using technology to study situations,
collecting as much data as possible on the basis of understanding things
and finding solutions that we are not able to reach with the tools available to the human mind.
How can these deep learning programs achieve such results?
Very simply, they are able to analyze an infinitely higher amount of data
than we could ever do, not only using our mind,
but also programming us software to do it in our place.
And then among all this data, simply for a matter of money,
they can find correlations that we would never even have dreamed of looking for.
For example, as Kai Fuli refers in his bestseller on artificial intelligence,
the algorithm discovered that the propensity to return a loan to the bank
correlated with the fact of having a certain average level of load on the client's smartphone,
with a certain average speed with which the client inserts his date of birth in the form
and with the fact of signing the loan documents on Wednesday.
How do these correlations explain themselves? We have no idea.
Nor would it ever occur to us to try to see if there are any, because for us they make no sense,
because the fact of asking for a loan on Wednesday should make us more reliable than those who ask for it on Tuesday.
We don't know, but artificial intelligence has found them.
And so, once we realize that the algorithm is always perfectly right in making the right decisions,
regardless of how absurd their methods and parameters may seem to us,
we could do nothing but raise our arms and say, ok, we give up, you study the issue and you decide.
So, for example, to choose the most suitable candidate for a job,
choosing a job interview today takes into account the study titles,
the skills, the emotional intelligence, the relational skills, etc.
Tomorrow, in his place, there will be an algorithm that will perhaps discard a candidate
because he is dressed in blue being a Tuesday, he was born in November,
he prefers blueberries instead of raspberries,
and on Monday he put like to a certain TV series on Netflix.
All things that seem completely absurd to us, the problem is that he will take us.
In the sense that these parameters, which seem to us to cry out for revenge,
will be exactly the right ones to predict perfectly if the candidate is suitable or not for that job,
not only to do the job perfectly, but also to feel happy and fulfilled and paid for doing it.
And here, in the example, we mentioned parameters that, if nothing else, we can also understand,
even if we don't explain the correlation,
but the algorithm will also have thousands of others to which we will not even be able to give a label.
Ok, so these are the terms of the question.
Like all new things, it poses its problems, its limitations, its risks, its moral dilemmas, ethics, regulations,
especially as regards privacy intrusions,
all things that we will obviously have to study carefully and solve in the best way possible,
but, in the end, it is a technological revolution so crazy in terms of potentiality
that, if used in the right way, there will always be more advantages than disadvantages,
since progress can be slowed down, but not stopped.
And so, sooner or later, we will have to make sense of it.
The main dilemmas are not so much the danger of the science fiction novel
that robots become masters of the bad world and exterminate us all,
but simply the fact that they will have so much information available on everything and on everyone
that nothing will escape, no one will be able to do anything without a computer knowing it.
And so this obviously poses control problems
in the event that this information is either simply used by private companies
that, if they manage to combine their profits with the general well-being, do so willingly,
but, when the two things clash, they will always choose profit over general interest.
Or, even more serious, if these technologies come together in the form of a dictatorial government
that makes arbitrary and contrary decisions in the general interest of the citizens,
justifying them with alleged emergencies of security or public health or financial stability or other.
On the other hand, the positive potentials are exterminated.
Artificial intelligence will be able to diagnose diseases in a perfect and very early way,
including those more rare or exotic that often bone-in doctors cannot recognize and diagnose.
But not only to diagnose, it will also be able to make better medical decisions.
Again, the human mind of a doctor, even the best one,
becomes pale in front of the ability of a computer that knows everything,
can process millions of variables, compare with the previous cases from all over the world,
knows all the drugs, their interactions, the side effects, all the surgical options, etc.
And in addition to making perfect medical decisions, it will also be able to notice potential problems,
to tell us three days in advance if we are about to have a heart attack,
to understand if someone somewhere in a dark and deserted street
has felt unwell all of a sudden and has no way to ask for help.
And these were all examples of medical applications,
but we could make analogies in the field of education and training.
In the field of agriculture, in the field of meteorology, in the field of security,
there would certainly be many less crimes,
because what robber, what thief, what homicide would have the hope of making it frank?
You can also make a fool of the marshal, but the algorithm that knows everything, no.
The more he knows everything about everyone, the more he understands everything and finds effective solutions.
The main problem of these softwares today is not to program them, but to give them data,
as much data as possible on everything, because the more data they have, the more they understand things and find solutions.
And this is basically the main factor that slows down artificial intelligence today,
not technology, not programming, not materials, but access to data.
And that's why China is in pole position, because of course the United States are now slightly ahead in technology,
but that is a gap that will be easily filled.
The real problem will then be having the data to feed that technology.
And no one beats the Chinese, because they have very few bureaucratic links of data collection and use.
All this combined with their mega-apps that do everything, from payments to the purchase of cinema tickets,
to the list of shopping, the booking of medical visits, the shipment of referrals.
So the same app knows what you spend, what food you order, how much gymnastics you do, the results of your blood tests, etc.
In the United States, however, there are different apps that deal with different things, so no one really knows everything, right?
Amazon knows what books you order, but not what you eat, what you spend and what your health conditions are.
And vice versa, Facebook knows what you think about many political, moral and ethical issues,
but not what you spent. The supermarket app knows what you spend, but not what you eat at the restaurant or the food you get delivered at home.
And it doesn't know anything about your clinical history, so it can't correlate the type of food you usually buy with your health status, etc.
In Europe, then, we are the most disadvantaged from this point of view, because in addition to this fragmentation of data,
we also have an almost devastating excess of regulation by the European Commission on the treatment of data,
to the point that any collection of personal data of users becomes a legal nightmare.
Not that privacy protection is not important, of course, but we must find reasonable compromises,
or at least make it clear to everyone that if we exceed on one side, we lose a lot of competitive advantages on the other.
Because obviously a citizen, you tell him, do you want your privacy to be protected? He says yes, of course.
But you also have to understand what you are missing, because when will the Chinese and the Americans have the technology to diagnose and cure any disease perfectly?
Give quick alerts in case of accidents?
Predict in a flawless way earthquakes, earthquakes, volcanic eruptions?
Teach any subject in the most effective and well-trained way to each student,
based on what he already knows and what he has not understood, and the way his brain works?
Predict the weather wonderfully, the best times for sowing and harvesting,
the most suitable measures to take from time to time to save the harvest,
select the most suitable workers for the various tasks by zeroing all incidents at work,
predict in advance when one is about to commit a violent crime,
manage traffic by zeroing traffic jams and always releasing parking lots where and when needed,
and thousands of other incredible applications in every field to improve the lives of citizens,
obviously we will also want all those advantages, but we will be completely at the mercy of their technology,
because here we had our hands tied every time we tried to collect some data on a large scale to train an algorithm.
So the main problem today is data and legislation rather than technology.
Technology is already very powerful, but we already know how to collect data wonderfully.
You will surely have noticed that if you repeat a certain number of times a word in a conversation with your smartphone nearby,
even off, off in reality you never turn it off, it is only in stand-by with the screen off,
but it works, maybe you are at lunch with a friend with the phone on the table and you repeat several times a word in the conversation,
I don't know, Portogallo, ah Gigi has then gone to Portogallo, yes, because in Portogallo now there is still time, etc.,
then you open the cell phone and do a search and you start to see the advertisements,
extraordinary offer, booked now, hotel in Portogallo from 76 euros per night.
It always listens to you and you actually allow it because if you want it to turn on when you say
Hey Siri or Hey Google Alexa and turns on to carry out your orders,
but it is evident that in order to hear when you say Hey Siri or Hey Google Alexa,
it must also listen to you all the rest of the time, no, in continuation,
understand what you say and turn on the screen if you say that word.
Meanwhile, however, it also hears everything else and takes the opportunity to give you some advice for purchases.
Now, today all this happens there, without ever leaving your device,
it is encrypted, so no human recording can even hear that you are talking about Portogallo,
there is a business between you and your smartphone, tablet, computer, smart TV, Echo, HomePod, etc.
But it is only for a legal link, which is the day it will be possible to do it,
this encryption is removed in half a second with a click and at that point
artificial intelligence programs will be able to start working with billions of billions of data
collected in real time 24 hours a day from anyone who has a smartphone nearby, that is, everyone.
And so we come to our field of interest, the nutritional one.
Do we see the potential of this technology even in the field of nutrition?
Obviously yes, I see huge potentials and I was telling you in particular three.
The first, probably more feasible in reasonable times,
will be to help us understand what people eat.
It seems trivial, but as we discussed throughout the video,
tell me what you eat, within the series on the limits of nutritional research,
understanding what people eat is actually one of the major problems we face in the field of nutrition,
both at the research level and in the clinical field, that is with patients.
Research in the sense that when we do epidemiological studies,
to say for example, ah, we have seen that in a population of 50,000 subjects,
those who eat less selenium have more incidence of cardiovascular diseases,
Here, but understanding what people eat is very difficult.
Even today, almost all epidemiological studies are based on the submission of questionnaires,
or those based on memory, or those based on frequency estimation,
or the food diary, which then an operator must translate into numerical data,
trying to find out how many calories, how many saturated lipids,
how many omega 3, how much vitamin B1, how much zinc, etc.,
is in that person's diet.
And obviously all this process is on the brink of huge mistakes.
It is the same when a nutritionist, in front of a patient,
the most important information to be able to help him improve his diet
would be to know precisely what he eats, how much he eats, and how he eats.
But it is an equally difficult task, because in a one-hour interview,
he has to ask all the questions to try to understand it from scratch.
Even Nipo asks him to keep a diary, use questionnaires,
and then do the same work of guesstimate, that is, estimate with a little imagination,
to try to get an idea of his diet.
Today, the best way is to ask him for a week to photograph everything he eats,
describing each photo in as much detail as possible,
hoping that week is indicative of his habits,
hoping he doesn't forget anything,
and then he will have to study all the material to try to guess the diet.
But it is a long, laborious and tedious operation,
both for the patient and for the nutritionist.
So the truth is that today, with the tools we have available,
it is really very difficult to understand what a person eats,
unless we live 24 hours a day for several months,
which is obviously not possible for a nutritionist with his patient,
and much less for an epidemiological study with tens of thousands of subjects.
What can be done instead with the potential of artificial intelligence
will be to develop image interpretation applications,
like the ones we already use today to identify a plant by taking a picture of it,
but doing the same with the dishes,
that is, simply photographing what one eats,
the algorithm will be able to deduce what it is, what the ingredients are,
what the quantity is, and then instantly translate everything
into an estimate of the number of calories, macronutrients, vitamins, minerals, etc.,
contained in that dish.
Obviously, going from the photo of a plant to the name of the plant is much easier,
going from the photo of a dish to what the ingredients and quantities are
is infinitely more complex, but not impossible,
because the algorithm will be able to cross the data with the habits of that person,
those of the place where it is,
if I take a picture in a Korean restaurant or inside a McDonald's,
those of the other information collected in that geographic area
where the ingredients and habits are more or less homogeneous,
maybe if it will be possible also by crossing the data
with what I bought at the supermarket or ordered on Glovo.
And since the glasses or the watch will immediately notice when you are chewing,
it will no longer be possible to forget to record the data,
because they will remind you in real time to take the picture,
if you have smart glasses, they will do it directly without even having to warn you.
In short, I think, given the exponential speed with which this technology advances,
that we could get there in a decade,
and so our work will be infinitely easier,
because when a new patient arrives,
we will immediately have a very precise and detailed idea
of his eating habits,
and therefore we will be able to give him advice and the most relevant and targeted instructions
for his specific situation.
The next challenge,
if we get to a greater integration of data,
and therefore it will be possible to cross the data of our eating habits
with the data related to our health,
our biometric data, our clinical history,
we will finally be able to solve many of the controversies
that today are still far from being clarified
and discover a potentially infinite number of new associations,
correlations and causal relationships,
incredibly precise effect between what one eats
and the effects, both in the short and long term,
on his health and on the one he gets sick of.
And we will also have all possible confusing factors under perfect control,
because we will also know his level of physical activity,
the quality of his sleep, his level of stress,
how much he smokes, how much alcohol he takes,
his income, his education, his profession,
social relations, etc.
In practice, a constant epidemiological study
with all possible parameters under control
and the entire world population as participants,
followed for life.
Imagine how many crazy things we will finally be able to discover
beyond any reasonable doubt.
Is it better to make one, three or five meals a day?
Is it appropriate to take a multivitamin every day?
Does adult milk do good or bad?
Is it a problem of milk or of the hormones and antibiotics it contains?
Do vegans or vegetarians who eat eggs
or those who do not eat fish eggs eat better in the long term?
Or do breads or steaks eat better for breakfast?
Is the consumption of red meat always associated with an increase in colon cancer
or only with certain types of meat,
or only those raised intensively,
or only if we don't eat enough vegetables together,
or only in those who do less than half an hour of physical activity a day?
When we will also have crossed the data of genetic sequencing,
which is becoming faster and cheaper,
so soon we will be able to sequence all our DNA
with the same ease with which we measure glycemia today,
at that point we will also be able to understand
if there are actually relationships with food
in order to be able to say who carries this and that gene
is better off with a ketogenic diet,
who this other with a vegetarian diet,
or maybe not, and I think not,
I think we will discover that there is no relationship
between genetics and food advice
and that we will conclude that the optimal diet is one and the same for everyone,
but finally we will be able to know for sure,
today we can only make hypotheses.
All the epidemiological studies we do today,
even the largest, with hundreds of thousands of people
who are followed for years and years,
are always very very limited,
they detect the habits by questionnaires
and then they choose a predefined number of parameters
to monitor, maybe even tens,
but still a finite number,
in general which nutrients one introduces,
what eating habits he has, how much he smokes,
how much physical activity he does, his work, his income, his education,
but certainly they do not detect thousands of other things
that seem not relevant,
like if he takes classical music lessons
or if he buys all his clothes from a certain brand,
and they can not even measure rare events,
on a study of 100,000 people,
which is still huge,
it makes sense to see how many have a heart attack or a tumor,
because even a rare disease that affects a person out of a million
would not have any statistical value,
but not even a less rare disease
that still affects one out of 20,000,
because in a study of 100,000 people
you will have 5 cases, and what correlations do you want to find on 5 cases?
For those, today, we have to do retrospective studies,
case control, which are even weaker.
Artificial intelligence, on the other hand,
will have the data of billions of people available
and it will know everything about them,
everything about everything and in an incredibly precise way,
and at that point the correlations will become statistical certainties,
then scientists will be able to have fun
if they want to investigate the mechanisms
to discover the causes of those correlations,
but it will have become almost a superfluous pastime,
because there will be no longer any doubt
on the certainty of the association
and on the relationship of cause and effect.
Unlike today,
when we do an epidemiological study,
then the statistics on the data,
even if they are statistically significant,
we must always say
that association does not imply causality,
we cannot conclude anything
until we confirm these data
also with a clinical study
and until we are able to give a biochemical explanation.
Here, all this process
will become useless from the practical point of view,
for the simple reason
that it will be the simple force of the data,
of the immense spring of billions and billions of data,
to put the green light
to guarantee that there is not even the shadow
of a doubt
on the certainty of the conclusions of the algorithm.
Ok, so this will be
the even greater potentiality,
that is to revolutionize the way of doing research
and provide much more precise answers
to questions that today,
with the type of scientific research
that we can do today,
we cannot even hope
to get closer to an answer.
This type of approach to knowledge
in which the data will be so many
to make superfluous the old method
of inference for induction and deduction
is called dataism,
that is, when the spring of data
reaches a certain level,
there is no need to understand them anymore,
because they will really speak for themselves,
something that today they never do.
There, instead, if the algorithm tells us
who buys Armani's clothes
is more at risk of Crohn's disease,
maybe we will not have the clearest idea of why,
but we will be sure that it is so.
And let's move on to the third potentiality,
that is to help people
to modify their behavior
and their lifestyle
by means of very effective
and perfectly personalized strategies
of education,
teaching,
persuasion and motivation.
You may remember that in the video
on the personalization of the diet
we talked about how much
strategies can be personalized
on the basis of our interlocutors
to persuade them in the most effective way,
taking advantage of their way of thinking,
their emotions, their values,
their fears,
we can adapt
the way of communicating
in order to persuade them more easily
and induce them to certain actions,
and this by personalizing
the type of messages to reach them,
their framing, that is to formulate them
in the most effective way for that person,
make them arrive at the right moment
in which they will be more receptive,
use the words that make them more reluctant,
make an emphasis on the concepts
that make more impression on them,
maybe knowing that a certain thing will make them indignant
so it will be easier to induce them
to act on something, etc.
This is actually the easiest goal
of all to the current state of knowledge
and technology, but what I think
we should choose to focus more
for the simple reason that
when you use such powerful tools
to make people do something,
you have to be 99.9% sure
that that thing
is useful, necessary
and positive, and today
we don't have this type of certainty
for absolutely nothing,
so first we should wait
for the two goals we talked about
to be realized, and at that point
we will be so sure
of which are the best choices to make
that we could also switch to the practical application
and, let's say, inculcate them
to the population.
Doing it today, instead, would mean
taking the risk of taking
really colossal shortcuts.
I mean, if they told me, look, we have the perfect app
that will induce all people
to follow exactly the diet you say
to eat what you say, when you say it
and in the quantities you say
you just have to tell us what this diet is
and then we make it followed all over the planet
well, I would give it to myself and I wouldn't
feel like doing such a thing
because there are still too many things we don't know.
So, I repeat, at this point
I think we should be very careful
and decide to put
in stand-by this type of applications.
And I repeat,
by choice, because already today we have
tons of information
on how to persuade, manipulate
and induce people to do things
feeling like they were the ones
to decide them. We have so much
research in the field of behavior modification
born in the field of
social psychology, developed a lot
by marketing, which then exploded
in the field of intelligence
imagine what we could do
when this type of knowledge
will be put in action not by a group
of human advertisers or strategists
of an electoral campaign
or an artificial intelligence program
that, in addition to applying these
knowledge, will be able to study us
and accumulate it exponentially more
to become incredibly
able to manipulate us and make us
do exactly what we want without
even realizing it. Those algorithms
will surpass in a minute long
all our current knowledge
in the field, both public and
still covered by the military secret.
Knowing a lot about you,
what you like, which movies you watch,
what you like on social media,
which channels you follow, which links you click
and which ones you don't, how you
answer a chatbot's questions
they will know exactly what
attracts your attention
the type of images that
surprise you, the way to formulate
the sentences to make more emphasis on you
the things that send you crazy, how
will motivate you to take action. Because knowing
what you like, what you want, what are
your values, your interests, the algorithms
will often understand more things about you
than you know yourself.
You will remember a few years ago
the case that ended on the newspaper of that girl
to whom Google continued to offer
cool ads and she said, what do I care
about the cool? And then she was pregnant, only
that Google had noticed it before her, from the
research she was doing, presumably
on the symptoms, I don't know, like what to take
for nausea, etc. Maybe a few
weeks before she had assigned a gallant appointment
on the calendar. Or
many people decide for those who vote
a few hours before the elections, but
their computer already knows much before them
if they will vote and what they will vote.
In some cases, as you will remember
with the Cambridge Analytica scandal, it was
the algorithm that used
the profile of the users to
calibrate the strategic communication
and induce them to vote
in a certain way. So
the result of all this will be that
artificial intelligence programs
will become incredibly good
at manipulating us and making us
do what they want without
being warned. And this is the key
for those of us who always remember, as we have
already said many times, that
manipulation means this. Today
many people feel like victims of
great manipulations on extremely
coarse and divisive issues,
but these people are wrong
for a simple reason, that is, that already today
we know how to do much, much better.
When there are obligations, prohibitions,
impositions, there is no great plot
or manipulative design, or if there is
you are not effective, because it is very
negligible. It is not an effective
manipulation to hammer every day
saying, do this, do this, do this, because
there one says, but why do you make me do
at all costs this stuff, what is behind it?
There will always be thousands
of people who shout and
oppose each other and therefore millions of other people
will remain at least in doubt
to say, who knows this thing,
if I can trust it? No,
effective manipulation is when one
does not realize it and there is no need
to force, nor to forbid, nor
to try at all costs to convince.
It is not to make people do things
in spite of themselves, but it is to make them choose
to do them and they are also
very happy to have made that choice.
And this happens when
imperceptibly, day by day
in the most subtle ways, one moves
tick, a little piece of brain,
then tick, tick, tick, and day after day
our perspective changes
and our way of seeing the world and thinking
things until we decide
in a certain way, but we will decide
of our initiative, without anyone having to
push us. And this is done
in a subtle way, speaking in appearance
of other things, making us arrive
little pieces of information, so
impassant, so that then everything happens
in our brain and we will be
ourselves, without even knowing why,
to do certain things,
make certain decisions,
have certain ideas, values,
appreciations, etc., and obviously
also to decide what to eat.
For this reason, it makes me smile that there are
those who think that the great design of manipulation
of the industry or the elite is to make us
eat the flour of the grills,
that so much no one will ever eat in Italy
except some cool Milanese
son of the notary who spends 40 euros
for a grill hamburger in the fashionable place.
Other than the grills, the crickets
that we are, that we don't notice
when they really manipulate us
and when, instead, they use only
caricatural diversions to distract
our attention from real manipulations.
Other than insects or synthetic meats,
let's look at what already today
fills 95% of the shelves
of supermarkets and that we buy
happily, because the problem is there,
in what we already do voluntarily,
not in what we will never do.
An example of manual manipulation,
for example, in Italy, was the one
that led two-thirds of the Italian
population to buy bottled water
for home consumption,
making Italians the first consumers
in Europe of bottled water,
despite the fact that the quality of our
water, you will say, is among the best
ever. In other words, the industry
has convinced millions of Italians to spend
a lot of money for something
not only completely useless, but even
worse than the almost free alternative
which is tap water.
But no one forces them, you see,
they choose to buy it because brainwashing
was so quick and effective
that they are convinced that bottled water
is healthier than tap water,
but not because someone told them
these terms, no? Because if there was
the minister on television saying
know that bottled water is better than
tap water, open your eyes, everyone
would have immediately said, ah, who knows
what's under it, who knows what it earns,
it will be a sale of multinationals.
The next day, twenty Facebook groups
would have come out to show all the risks
of bottled water with tens of thousands
of users who would have said, ah, you see,
they want to manipulate us. Eh, this is not
the effective way to manipulate if the
manipulated one realizes he is. Instead,
bottled water is made through crossroads,
bringing a little piece of information
from here, one from there, one from above, one from below,
the leaflet on the newspaper with periodic
cadence that says, ah, the water of the municipality
has exceeded for three days the limits of this
and that, but rest assured that you can drink
the same, so that then the click
happens directly in the heads
of the people who will be the ones to say,
ah, wait a moment, but then it's better
if I buy bottled water. But they will have
the feeling that they were the ones
to decide it, not that someone imposed it on them.
And at that point, even if someone
had to doubt their choice,
they would be the ones to defend it
convincingly. They will say, eh,
but who knows, the pipes in my house are old,
but anything that can migrate
in the water that passes in a pipe with which
it is in contact for ten seconds will always be
less than what they take in the plastic bottle
in which it has been for months. If then
the light of the sun is exposed to a source of heat,
let's imagine, it is filled with microplastics.
Eh, but that of the aqueduct is contaminated,
but that of the bottles can be more
because mineral waters have limits
much less constricting than those of the aqueduct.
Half of the mineral waters
on the market, if they came out of the tap,
the mayor should immediately issue an order
to stop drinking it, because they are within
the limits set for mineral waters,
but they exceed those set for
tap water. And this is the analysis of the source,
but we don't drink the source, we drink it
after it has been in a plastic bottle
for months. For decades, that bottle
was made of PVC, which then
was discovered to release in the water
carcinogen. And so, after millions
of Italians have drunk
cancer for years and years, from one day to the next
we have simply said, oops,
no, sorry, the PVC was carcinogen, but
rest assured that today it is forbidden, now
we make bottles of PET. And all went well,
then maybe in another 20 years we will discover
that even in PET there is something carcinogen,
and amen, we will make something else of it,
and in any case millions of Italians will continue
to buy their bottled water, because they will remain
convinced that it is still healthier
than that of the tap. Ok, here it is, but
this manipulation came so well
more by a lucky coincidence,
lucky for the multinationals, not for us,
but by coincidence, imagine
if this type of manipulation
could be done scientifically
and systematically for everything
you want. That's why you understand
that here we are in a minefield,
because already with the tools available
today we have done not a few damage.
Imagine what we could do
if we give these tools
to the artificial intelligence
that will be able to elevate them to power
in an instant. Before using these tools
on a large scale, therefore,
it will be necessary not only to have the certainty
much more than reasonable, I would say almost absolute
of the things we want to induce
people to do, and today I would say
that we do not have this type of security
in any field, nutrition included,
but it will also be necessary to establish
a solid mechanism of weights and counterweights
so that the control of these tools
always falls in the last analysis
in the hands of all citizens
and not of exclusive domain
or private companies
and much less of individual governments
or other narrow circles of
alleged enlightened people who can decide
from their initiative how to manipulate
the population, maybe even in good faith
but with the risk of taking
crazy stunts.
Unconditional supporters of artificial intelligence
say that this is a problem
that will no longer be solved because
when the artificial intelligence itself
will know everything about everyone
we will not have to decide
how to manipulate the population
for the better, but it will be decided
directly by the algorithms.
The problem is that at that point
we will be really out of the loop
and above all that making decisions
only on the basis of data,
which is the only thing an algorithm
can ever do, is not the most
shareable choice in terms of ethics
or morals, which as you know
are not universal, but vary
depending on the historical period
and the geographical location.
And so these are dilemmas that a robot
risks solving in a very different way
from how we humans would like.
And with this, gentlemen,
we conclude our look
in the crystal sphere of the future
which, as everyone knows, is always very
tricky because then things can happen
completely unexpected and unpredictable
that combine all the cards
on the table from one day to the next
changing all our predictions
and our way of seeing things
in a way that no one could have
foreseen or imagined, but in short
to imagine what the future
could hold for us is still
a serious exercise of exciting
and stimulating prospects and above all
it is not a sterile exercise
even where everything should change
in an unexpected way, because
if we want the future to take
a certain direction rather than another
we have to start thinking about it today.
And as for the present,
there is a rule to never forget
which is this one
we can also give it to the algorithm
quiet life, happy mind
and moderate diet.
All clear? Very good.
