Social Effects by the Singularity -Pre-Singularity Era-

Social Effects
by the Singularity
～ Pre-Singularity Era ～
Hiroshi Nakagawa

Reference
• Ray Kurzweil: The Singularity is Near ，Loretta Barrett
Books Inc.2005
• James Barrat：Our Final Invention: Artificial Intelligence
and the End of the Human Era, New York Journal of
Books, 2013
• Nick Bostrom: Superintelligence, Oxford University
Press. 2014
• John Markoff: Machines of Loving Grace: The Quest for
Common Ground Between Humans and Robots ,2015
• Thomas H. Davenport , Julia Kirby :Only Humans Need
Apply: Winners and Losers in the Age of Smart
Machines , Harper Business, 2016

What is the Post Human that comes
after the singularity?
• Kurzweil writes that the singularity is :
– Technologies about Ｇ（gene）、Ｎ（nanotech.）、Ｒ
（robot including AI） develop exponentially as
depicted in the right figure
– ＊ is singularity point. After it, tech. grows rapidly to
the infinity
• Kurzweil predicts the singularity will occur in
2045.
• Post human is a new version of human that is
transformed by G, N, and R after singularity
point.  Not a human anymore
time
performance
＊
Singularitypoint

Image of post human
Human 1.0
Today’s human
Nano-bots made from carbon
nano tube are injected into
blood stream, heal the disease
parts and communicate to
outside of body.
Gene
transformation
makes completely
immune from
disease, and even
immortal.
Knowledge is up or
down loaded between
the brain and outside
via internet.
Human2.0
or
Post･human
Immortal and acquire all the
knowledge all people have.

Pre-Singularity
• Is a post human really possible?
– Puerto Rico Conference on 2014 organized by computer
scientists said it is impossible.
• This conference is a counter part of Asilomar Conference of bio-
scientists.
– Especially, nano-bot is so hard to realize that Super Artificial
Intelligence( Super AI) which is much more intelligent is
necessary to develop nano-bot.
– Some researchers say that apoptosis should be implanted
to AI in order for Super AI not have proliferation （Barrat Ch.
14)
• Apoptosis is a process of programmed cell death that occurs
in multicellular organisms.

Pre-Singularity
– Even in Pre-Singularity era, meaning the time before
the singularity point, development of advanced AI will
causes various types of social, ethical , technological, or
even humanity related problems.
– The remaining part shows these problems and
hopefully the solution to these.
• Unfortunately, hope is very small.

Contents
• Stance of scientists community against Pre-Singularity problems
• Amplification vs. Replacement
• AI takes over jobs
• Boarder line between amplification and replacement
– Autonomous driver： trolley problem
– The right to be forgotten
• Towards black box
– Responsibility
– Vulnerability of financial dealing system made of many AI agent traders
connected via internet
• AI and weapon
• Filter bubble phenomena
• Analogy： Selfish gene
• AI and privacy
– The right to be forgotten, Profiling and Don’t Track
– Feeling of friendliness to android
• Again self conscious and identity

Stance of scientists community
• Asilomar Conference held by computer scientists in
2009 rejected the Kurzweil’s idea of self autonomous
AI realizes. （Markoff Ch.10）
•
Google had acquired Deep Mind in 2014, then
computer scientists organized Puerto Rico Conference
on 2014, but they were extremely passive compared to
Asilomar of Bio-science.（Markoff Ch.10）
• Many people think it problem that computer scientist’s
community does not have a sense of crisis about AI.

Amplify vs. Replace
• Does Artificial Intelligence amplify human competence?
– ＩＡ: Intelligent Assistance/ Intelligence Amplifier
• Does Artificial Intelligence replace human?
– ＡＩ: Artificial Intelligence
• “IA vs AI” is a basic view of Markoff’s book: Chapter 4 beautifully
describes this view and the history of AI.
• IA vs. AI : This is a 60 year old hostile or complementary
relation since the starting time of AI.
– When AI is high, IA is low, or vice versa.
– If AI technology becomes stuck, researchers become favorite IA.
– Question: Is Deep Learning IA or AI?

Is a human in the loop or out of the loop?
• Designing principle parallel with “IA vs. AI” is “a human is in
the loop or out of the loop .”
• In the loop  IA
– A system is an extension of human abilities. A human being and
a system do work collaboratively.
– A human does some task being aided by an AI based system.
– A human might not understand what is going on in an AI system.
• Out of the loop  AI
– A system acts autonomously. A human only commands a system.
– A human is no more an actual stakeholder.
– A human is usually living in a very easy position, but could not get
by when some thing happens.

Is a human in the loop or out of the loop?
• In the loop  IA
• Out of the loop  AI
• Generally speaking, the problem is to find the criteria
of how far AI system should or can be done on behalf of
human beings.
– Of course, the remaining is to be done by a human being.
• This is a traditional problem between machine and
human, but becomes complicated in the era of AI.
• To leave all the decision-making to AI is too easy for
a human being but ends up with a kind of addiction
and intelligently atrophic(shrinking) .

Example : Self-driving cars
• self-driving: Level 1,2,3,4,5
– Fully self-driving (no human intervention): Level 4，5
– Human intervention is required at a critical moment: Level 3
• If you usually don’t drive and are handed over driver’s
position at the critical moment, can you drive better
than AI based self-driving?  NO!
• Thus, Level 3 is much more dangerous than Level 4,5
especially at the moment of taking over, isn’t it?

AI takes over human jobs
• It is not new that new technologies take over jobs
previously done by human, such as industrial
revolution.
– It is too reasonable because technologies have been
developed to replace the heavy burden of human jobs or
to improve efficiency of jobs done by human.
– Fortunately, when one type of job is taken over, other
jobs come in, and many people found new jobs.
– Many people optimistically think the situation is the
same as AI, but is it really so?

AI takes over human jobs
– Many people optimistically think the situation is the
same as AI, but is it really so?
– Reasons for pessimistic view:
• General, or Strong AI has not only high performance for
specific field such as IGO but also great flexibility for
various , even new, fields or tasks.
• This kind of flexibility and versatility of AI will be
strengthened rapidly even in Pre-singularity era.
As the results, various types of jobs are expected to
be taken over by AI.

 Various types of jobs are expected to be taken over
by AI.
 Researchers who are thought to be most distant
from this crisis.
 Published papers skyrocket up to millions a year. Then
researchers are facing though time to cover all of them
even in his/her own or somewhat related area.
 AI reads and understands countless number of papers,
and extract enough small number of paper that are
strongly related to a researcher’s interest.
 This has already been AI addiction.

 AI reads and understands countless number of papers, and extract enough
small number of paper that are strongly related to a researcher’s interest.
 If papers are read by not human researchers but AI, what matter in
research activities becomes not writing papers but doing experiments
based on the proposed protocols and getting the results data.
 In addition, what we want AI to do is to find niche subjects or even
research topic never explored.
 At next stage, AI selects research topics, do experiments in a standard
procedure, get results and finally publish online the results. No
human researcher is needed.
 We human have to re-define the concept of researcher in AI
era.
 How to re-define?

Davenport says we can find new jobs
immune to AI invasion, but…
• Davenport shows how to find AI free jobs in his book:
– Jobs aiming at higher quality than AI, say judgment without
data  It must be done by so called genius?
– Jobs AI can not do such as human intercommunication,
persuasion, etc..
– Jobs of finding connection between business and technology
– Jobs less economical with employing machine or AI , in other
words, quite rare and case specific task.
– Jobs of developing AI
• However, AI itself will develop AI in the near future!
– Jobs of explaining the results generated by AI or action done
by AI.
• Again, however, it will be done by AI!

Davenport says we can find new jobs
immune to AI invasion, but…
• Davenport’s proposal of the ways to find new job
requires too high level talent and skills for ordinary
citizen.
• Thus his proposal does not work in ordinary people.
– Seemingly it’s a preach to ordinary people from a high
social position.
• So, how should we do when our jobs are taken over
by AI?
– We need expanding welfare, or so called Basic Income.
– A life of welfare recipient equals no motivation life. In
addition heavy burden of welfare budget even if we call
it “basic income.” （Davenport Ch.10）

How to cope with this situation?
• The future social system might be a caste system of ancient
India?
– We have to share the remaining small amount jobs: caste…
– Hopefully, at most free from inequality.
– Prof. Y. Matsuo propose to use AI for fair assigning individuals a job.
– He seems to think it as an ideal socialism.
 The problem of this AI based system is that no one is prohibited to
reject the assigned job by AI: if rejected, socially chaotic situation
could be happen.
 In Davenport’s book chapter 9, the quoted Hon Hai president ‘s opinion is
“human being is also living things. It is a tough job to manage millions of
living things.”
 From this reason, the society needs the taboo: AI never says the
wrong thing.

It’s not me but AI says so!
 the society needs the taboo: AI never says the wrong thing.
 Even worse situation such as:
 A superior gives a subordinate an absolute order because the order is an AI’s revelation,
or more softly based on calculated results by AI.
 A teacher forces the student to do some work because AI based learning software says
so.
 In these cases, no objection is permitted because AI says so!
 This would happen because a superior or a teacher abandon his own decision or
sabotages his essential task.
• Such a state that no objection is permitted to AI’s decision
might encompass the society without freedom of speech
nor even human rights.
 We need to design the society in which we have the right
to object AI’s decision.

If currently working technologies or systems are
automated based on AI,…
• If a human job is completely replaced by AI and a
human being is outside of job process,
• a human can’t take over the job at a critical
moment because he/she usually does not do it
because he/she does not have enough experience.
– It is usually said that a human operator takes over at a
critical moment, but obviously it does not work.
• Anyway, if a human being totally relies on AI in
actual job,…

• if a human being totally relies on AI in actual job, the bad
consequences are:
– Human workers are rapidly loosing their skills and could not
transfer their skills to the next generation people,
– then their skills will be forgotten forever.
– This means that once a certain type of job is taken over by AI,
we could not regain it forever.
• There are a few companies which had transferred its
factory to offshore due to cheaper labor cost of
developing country, again move back to domestic
because of less expensive automated production process,
but,
 automated production line does not need big number of
human workers, then the number of jobs do not increase
in total.

• In sum, autonomous AI does not necessarily make
people happier.
Loosing motivation  loosing skills  the skills are forgotten
forever.
Loosing Jobs will progressively increase generation by
generation.
Number of jobs created by technology advancement is not
big enough to compensate lost jobs.
• There is no clear consensus as to how autonomous AI
itself could have to improve own ability, even among
researchers community on AI.
Some researchers on AI don’t like to think of this problem
because it is an obstacle against technology advancement.

• But carefully thinking, we find many situation where AI is
useful to replace a human task such as:
 Diminishing workforce (by aging of society)
forces us to introduce autonomous AI robots.
A nursing-care robot for elderly people is useful or even needed.
A child rearing robot is much complex, but..
If there are a self-driving car which drops and picks up a child at a
nursery school or day-care center, it becomes a great relief for
working mother.
Door-to-door delivery to a remote area by an AI driving drone.
Automated scoring for mock exam and even pointing out weak
points of the examinee. ( In Japan, this job is known to be really
tough)

Boundary between amplification and
replacement
• It is rather difficult to identify whether a certain IT technology is an
amplification or replacement of human ability.
• Examples of apparent amplification:
– A power suits for an elderly nursing worker.
– Remote diagnosis by a physician.
– Augmented reality such as Google glass.
• Examples of apparent replacement
– Autonomous nursing robot
– Autonomous door-to-door delivery robot
– Autonomous scoring and pointing weak point system based on AI

replacement
• Ambiguous example
– An AI system which is usually working autonomously,
but sometimes requests us information: semi
interactive system.
– Example: Question answering system such as tax
return assistance
– Example: Recipe generation system which changes
recipes depending on the user’s taste and physical
condition, and the remained food in the refrigerator
– self-driving car of Level 3.

replacement
• Considering these examples, one candidate of boundary is the following:
– If a human user of the system can understand, and predict the system’s behavior ,
then it is an amplification,
– otherwise , say the system is a completely black box for the user, it is a
replacement.
• The examples of previous page
– An AI based Q/A system can’t complete the task by itself, and require the user
additional data, information or even instruction.
– The user thinks he controls the behavior of the AI system, but in reality the system
controls the user.
– While an AI system seems to behave autonomously, it abandons the task at critical
moment.
• are somewhere between amplification and replacement.
• Obviously, these example contains essential clue about how AI becomes
General AI.

replacement  how to cope with?
• So far, we only think of the solution by technology.
• But if a technological solution is not enough,
we have to think of legal system to prevent the
bad consequences.
– Designing insurance policy to cope with an accident
– Typically , the problem of self-driving car on the next
page.

How to act the critical situation while
driving automatically?
When a kid suddenly runs in front of Level 4, say AI driving car,
1. if it avoids collision with the kids, the car jumps into the
opposite lane and crash head-on with an approaching car
and the person on the car will die,
2. if it goes straight, it collides to the kid and he will be killed.
• What AI driving car should do in this case?
– AI driving car is a much safer( =lower accident rate) car than a
human driving car. Focusing too much on this type of case slows
down the development of AI self-driving technologies. “Ridiculous”
AI researchers might say.
– If we accept this AI researchers opinion, we have to work out the
way how to treat the case of accident in order to maintain our social
structure.

• One solution is an insurance policy for an AI based
self-driving car.
• The first problem for this type of insurance is who
takes responsibility at an accident.
– a car manufacturer, a dealer or a car owner?
– The study is going on.
– To make an insurance policy, we have to estimate the
financial value of person. This kind of topic was almost
taboo in AI research, but we could not ward it off
anymore.
– Actually, insurance companies have already done this
based on expected total income of the rest of life.

• At collision, does AI estimate both of the kid’s expected
total income and the driver( not driving, only the car
owner)’s expected total income, and make decision based
on this estimations?
– In other words, designing insurance policy does not give us the
solution essentially.
– AI researchers have to be keen to this kind of problem.
 Essential solution is to design the transportation system in
which this kind of accident situation never happens, such
as:
– The transportation management system in which human driving
cars and AI driving cars never share the same road at the same
period of time.
 AI can be a good candidate tool for the above designing.

Another case: The right to be forgotten
• The European Court of Justice rules in 2014 that an Internet search
engine must consider requests that the requester wants to remove
links to web pages where the information he wants to hide, based
on the right to be forgotten.
• More than a million requests of removing links has come to Google.
– To the best of my knowledge, it seems to be the balance between the
right to be forgotten and the right of free speech.
• Google employs 50 to 60 specialist to cope with the requests. Even
for Google, it is a heavy burden.
– The promising way is to employ machine learning technologies and use
many cases decided by human experts to build the classification
system.
– This classification system roughly classify the requests, and human
experts only look closely the boarder line cases.
– The difficult thing is to clarify the reason why the request is decided to
remove or not. The reason written in natural language sentences is to
be clearly articulated to the requester.

AI has already been a black box, so
what?
• No matter whether AI may be amplification or
replacement ,
– an AI developer has to take responsibility.
– But, we have already lapsed into the situation where the
stakeholder can’t grasp what is going on inside an AI.
– The stakeholders include：
• Developers of AI
• Human agents who offer source data which is used as supervised data
of machine learning.
• People such as working at retail shops or advertising for AI system.
• Users of AI.
 It is necessary now to make a legal system which clearly states
who is responsible.

Black box AI harmed financial business
 So called AI trader agent which is regarded as a black box AI , has
already been working in stock market etc.. AI traders are
connected via internet and the speed is one transaction per less
than millisecond.
 Trades are proceeding so fast that human trader can’t intervene.
 Under this circumstance, a certain type of trade may trigger the
bad chain reaction worldwide where human traders can’t
intervene in time.
 In May 2010, some AI trader predicted the default of Greek national bond
and took out a lot of selling order. As soon as the order was confirmed,
every AI trader felt a kind of crisis and started distress sales, ended up with
Dow Jones crashed by 1000 points instantaneously. (Barrat Ch.6 and Ch.8)
 It took a long time to identity the reason of the crash.

Connected AI systems
It is regarded as a set of optimal actions of many AI
agents connected via internet. The key point is that
they have the common language that is a stock price
and effects their behaviors decisively.
One AI system is not that powerful, but many AI systems
communicating each other may make unexpected and
uncontrollable situations happen as described.
 We, apparently, need the legal system that clarifies who is
responsible in unexpected situation such as above described.
 But, a legal system is necessary and sufficient?

We should not be optimistic
 AI researchers often discuss the ethics of AI from the viewpoint of
whether a stand alone AI behaves ethically.
 But, once many AI systems that are network connected start to act
autonomously or collaborate, we can’t manage nor control them.
 We have not yet found the way to cope with it. If we install a system to
cope with it, there surely are bad guys who try to find loophole for
money.
 That bad guy’s action is hard to be detected nor verified.
 Is it possible to develop another AI system who can detect and verify the
bad guy’s action?
 Another concern is : if an AI has the potential to learn the way to
understand and estimate the current situation and make a good decision,
AI is very near to having self-consciousness?
 Do AI researchers really do research as consciously considering this
concern?

Military and AI
• Using AI in military is inevitable because:
– AI has already been used in various ways such as reconnaissance drones
equipped with AI as well as planning or data processing in military.
– Basically AI technologies are open to public ending up with being caught
up easily by any countries.
– Different from nuclear weapons, AI technologies are expressed as
information. Thus they are easily and rapidly spreading via internet.
– In other words, terrorists, authoritarian state or enemy country are able
to get AI technologies to use for military purposes.
• In reality, Russia has gotten nuclear technology by using spies (Barrat Ch.14). In
today’s internet era, AI technologies are extremely easily spreading.
• The US military think if they are behind in AI technology, they become inferior to
other countries in military.
 As for AI based weapons, Russel’s opinion is important (Nature 28 May
2015, http://www.nature.com/news/robotics-ethics-of-artificial-
intelligence-1.17611#/russell )

Military and AI
• Some one says if you embed AI the ethical sense, it’s OK,
but,
– in the war time, if our AI robots act ethically, we will be
definitely defeated when enemy’s robots act unethically. So
embedding ethics does not work in the war time.
• A big portion of AI research funds has come from DARPA
historically.
– In the cold war time, DARPA gave a big amount of funds to
machine translation research in order to build an NLP system
which automatically understands huge volume of Russian
documents in short time.
– A developing of Siri is also funded by DARPA .
– AI technologies developed by DARPA fund are, of course,
transferred to use for military purposes.

Military and AI
• AI technologies developed by DARPA fund are, of course,
transferred to use for military purposes.
• But, the real problem is that DARPA might classify some
of them to keep superiority of the US.
– Nevertheless, the technologies are vulnerable to leak.
– Google that is doing research and development of AI secretly
( every body know this fact) hires DARPA high rank officers.
 Based on understanding these things, people including AI
researchers, of course, should discuss the dual use of AI.
Unfortunately, this kind of discussions are not so well
proceeding or sometime even biased.

What happens after AI takes over
human jobs?
• If AI takes over human jobs too much, people are
expected to be lose their skills or competences.
– Many persons become unable to travel nor even move in
city nearby without their smartphone.
– Many young workers only wait until the order coming
from computer.
– If a full self-driving car becomes widely used, car driving
insurance companies get severer damage.
• People who have driving skill would rapidly decrease.
• The accident rate of AI based self-driving car is much lower than
human driver.
• Finally people are not willing to get a driver license.

What happens after AI takes over
human jobs?
• If AI takes over human jobs too much, people are
expected to be lose their skills or competences.
– A (not highly skilled) lawyer is facing tough time!
– They become too much depend on an AI based search
engine for judicial precedents.
• Currently, the similarities used in search engine are calculated
by the vocabulary appeared on precedents or process verbal.
• If AI is able to investigate and analyze the background of cases
based on social common sense, no ordinary task are remained.
– The same situation is highly likely to happen on
accountants, tax accountants, workers at research and
development section.

Notorious effects on human individual
by AI: Filter Bubble
• Facebook uses AI to analyze and utilize its users’ profile
and information about his/her friends.
• The purpose is to recommend individual users the
information that seems to be favored by him/her.
• As the results, it becomes much harder for users to access
information that is not inferred as his/her favorite.
• Users are within the bubble made by AI system. The
information that AI decides users would not like is filtered
out and does not come in the inside of this bubble.
•  so called “filter bubble”

Filter bubble
• The filter bubble help develop the opposite to the original
idea if Internet : every one can access every information
on the world.
• Since users can’t access variety of information, they lose
diversity and flexibility of thinking pattern, say
informational degeneration.
• It is even harmful to encourage the bias or hate.
– People sometimes prefer this kind of information
degeneration because they don’t like to think things sincerely
and deeply with much effort.
– In addition, for IT service provider such as SNS, search engine,
it is much efficient and effective advertisement strategy only
to show the users’ favorite information.

Filter bubble
– However, showing only users’ favorite information goes
against the people’s right to know, and might lead to the
domination of information by a few people.
– Some IT company say that every thing is done by
software( including AI) and we didn’t intervene at all.
– For their viewpoint in terms of law, it is OK for them, but
is it really OK for healthy society?
We need an web tool which gives us the ability to easily
access the information we are not necessarily familiar
with nor favorite.
But, is it every thing? The profound cause of filter bubble
is in people’s psychology, or even human nature. No end
in sight….

AI and privacy
• The right to be forgotten.
– In the case that individual user sends the request of deleting link to his/her
individual information , the search engine company should decide whether
accept and delete the link or refuse.
– This decision depends on the balance between the right to be forgotten and
the right to know (or how important the search engine is publicly)
– The decision rule might be worked out by AI (machine learning) technology
with huge amount of decision historic data by legal specialists working at
the search engine company.
– Actually, this is a good and useful usage of current AI.
• Profiling
– Search engines or SNS build up their users’ profile with AI technology.
– User profiles are used in advertisement business which brings them huge
amount of profit.
– At the same time, individual user is very anxious about how their private
information will be used.
– AI creates a bright side and a dark site at the same time.

AI and privacy
• When an appearance of AI robot looks like human and
speaks as human being does,
– a user trusts the robot because a robot does not lie,
– feels an affinity because of its human like appearance,
• he/she tends to speak his/her own privacy to an AI robot.
• AI robots can collect personal or even privacy information
if they want.
– In order AI not to collect privacy information, the AI is to be
equipped with human common sense as to what is a privacy.
– What is privacy varies person by person, situation by situation.
– Going to restaurant with a lady is sometimes privacy for the
gentleman??.  AI with good common sense! Or general AI!

Misuse of AI
• An AI is easily expected to be used for some wrong or harmful purposes by a
bad guy. It is crime and punished by a law.
• Minor but implicit harm is( as stated before):
• A superior forces the subordinate to do a wrong or impossible task by saying “This is
what AI concludes the best or needed.”
• Maybe the superior does not have a logical reason, but just from his feeling or
sentiment. The subordinate gets suffered by a superior’s misuse of AI.
• It can be exaggerate but we need “veto pawer to AI’s decision (human
right?!?!).”
• At least, AI has to have a skill to explain the reason why the decision
the AI made.
• The issue is two fold:
– accountability of AI
– Trust for AI
• Both of them is difficult to implement at this moment.

Self-consciousness and identity
• AI is very near to take an action that is triggered by positive / negative
feeling to us human beings. If an AI acts like this, AI becomes General
AI in other words “singularity.”
• If an AI gains self-consciousness and identity, it becomes General AI.
 If an AI is designed to survive, the intension to survive is regarded as
its identify. The intension to survive requires to know the state of AI
itself. This is self-consciousness.
 Then, AI has a positive feeling to a person who wants AI to survive, a
negative feeling who don’t want AI to survive.
 As the results, AI acts according to this positive or negative feeling.
• This is a scenario Barrat and Markoff showed in their book.
• Unfortunately, they have not yet indicate the solution to this problem.

Self-identification
• Why , in the morning wake up, are we feel we are the
same person as we got asleep yesterday night?  mystery
of self-identification
• Kurzweil suggests: A human person gradually replace
his/her small part of body and brain, he/she still think
he/she is the same person, namely maintains self-identity.
• If this gradual replacing finally makes him/her a post
human, he/she(=human2.0) keeps self-identity with no
obstacles.
• Thus, a post human being has the same self-identity as
current human. A post human being will think as we,
human being, do. Really???

A tentative summary
• If AI is really needed to do a specific task such as elderly care, we have to
investigate more on AI technology useful in that task.
• A new AI technology is needed to explain or visualize for understanding AI’s
acts. For this, first of all, AI has to observe its internal behavior.
– i.e. AI should explain its decision of erasing link on the request of erasing the
requester.
– This observation and explanation mechanism is especially important in the field
of AI based trading where many AI traders are acting interactively and they are
connected via internet with stock prices as their common language.
• We need an AI which protects the problem caused by other AI’s
– Such as alleviating filter bubble effects.
• Of course, we need discussion if we adopt legal constraints.
– The examples of the same type is a nuclear weapon restriction, international
laws for war time, etc. . They have been conducted but there still is a long way to
go.

A tentative summary
• Considering the problems of this slide, before
saying exaggerate term “singularity”, we have
many things to think and try to find the
solution.
• Then, after singularity, what will happen?
• A part of answer is written in Nick Bostrom’ s
book “Superintelligence.”

Social Effects by the Singularity -Pre-Singularity Era-

More Related Content

What's hot (20)

Similar to Social Effects by the Singularity -Pre-Singularity Era- (20)

More from Hiroshi Nakagawa (20)

Recently uploaded (20)

Social Effects by the Singularity -Pre-Singularity Era-