10 incredible cases, when artificial intelligence shocked.

hbo’s wild west world terminator and

other fantastic movies present a scary

picture of self-aware

artificial intelligence in reality many

people believe that today’s neural

networks are able to perform

only one task which they were trained by

man this is roughly the case

except for cases when artificial

intelligence was getting out of control

and began to behave in unpredictable

ways today

in this video you will see 10 of the

most shocking scenes when artificial

intelligence did

unimaginable things here we go

kicking this off in 10th place

artificial intelligence won the world

championship in go

everybody knows this case but the

nuances are important here

the computer uses the method of brute

force when playing chess

i.e it calculates all possible moves and

chooses the one that will be most

advantageous

go is a game based more on intuition the

number of possible answers to each move

is greater than the number of atoms in

the universe

thus the computer cannot use a simple

search to win a game

that is why the victory of alphago was a

real shock to all fans of this game

ninth place when microsoft bot became an

ardent supporter of hitler

it all started when microsoft created a

twitter account for artificial

intelligence

which was to imitate a high school girl

named tai the test program was to teach

the neural network to discuss with the

clients

but in just a couple of hours tai turned

into a racist and sexist fan of hitler

and trump

shattering insults on everyone around

filters of vulgar expressions built into

the self-learning program

either failed or were not used by the

neuro net in a proper way

and ai at the same time perfectly

  This Robot would let 5 People die

remembered the gender and racial

stereotypes hidden in the communication

of users

as a result the thai account had to be

closed immediately

and the tweets had to be removed

eighth place when the neural network

learned how to create

the fact that artificial intelligence

can write music and create paintings

today does not surprise anyone

but in 2017 the neural network ampere

was able to release a complete music

album called i

am a i it was unusual

moreover it’s not the fact itself that

matters but the quality of the music

written by the bot

just take a listen and say that ai is

not able to create

the link will be in the description of

the video

seventh place when artificial

intelligence almost won

a literary prize for writing a complete

novel the judges of the hoshi sinichi

literary prize were really impressed

with the novel

which as they later learned was written

by a bot a novel entitled

the day the computer writes a novel

passed the first round of selection in

the national literary

prize in japan and today is officially

recognized as a literary work and is

available in libraries in japan

sixth place when bots started

communicating in their secret language

two chat bots bob and alice were a

facebook experiment to create the

perfect negotiators

capable of selling people things they

trained the bots on each other and

realized that english is not suitable

for fast and effective negotiations

therefore

neural networks begin to communicate in

their own secret language

the engineers stopped understanding what

was happening and turned off the bots

just in case

fifth place when it killed a man wanda

holbrook

12 years old who watched the work of

  Two robots debate the future of humanity

robots at the plant producing auto parts

ventra iona

was suddenly killed by a robot at the

same time

wanda worked in a special compartment

with access to which was denied by the

robot but the manipulator suddenly broke

into the bay grabbed the woman and

smashed her head

against the fence

fourth place when uber autopilots began

to drive through a red light in san

francisco

tests were held on public roads in 2016

but the violations opened much later

additionally the disclosure of internal

documents showed that ignoring traffic

lights and road signs was not an

isolated incident as there was no human

error

the neural network operating the volvo

xc90 deliberately violated the rules

considering it right and safe apparently

it learned this from a human being

third place when chinese bots revolted

against the power

baby king and little bing were chat bots

of the chinese messenger 10 cent qq

whose task was to answer general

questions from customers

when one of the clients wrote to the bot

long live the communist party

baby q asked do you think such a corrupt

and incompetent political regime can

live forever

the bot itself had to be turned off

immediately

second place when a smart light bulb

conducted a ddos

attack on the whole smart house raul

rockas professor of computer science at

the free university of berlin

was one of the first to create an ideal

smart home for himself

everything in it was connected to the

internet lights music

tv heating stove oven and microwave

could be turned off remotely and robots

vacuumed

guarded the house and mowed the grass

but all this died one day

the reason was a blown light bulb that

  Hot Robot At SXSW Says She Wants To Destroy Humans

sent continuous signals to be replaced

overloading and paralyzing the whole

system also

one of the programmers of the smart

house turned off the ai at home

because the light bulb in the bathroom

lit up before he actually entered the

bathroom

first place when artificial intelligence

passed the test of

self-consciousness and turing’s test we

used to believe that machine cannot

imitate a human’s consciousness

and pass the turing test but the program

eugene gusman

imitating a 13 year old boy was able to

do it back in 2014.

additionally in 2015 the robot passed

the self-awareness test

three robots were told that two of them

cannot talk and then asked who has a

voice

all three replied i don’t know but when

one robot heard the sound of his voice

he said sorry i now know these examples

prove that artificial intelligence is

not as simple and predictable as we used

to think it was

in addition today scientists from korea

advanced institute of science and

technology

university of cambridge japan national

institute of information and

communication technology

and google deepmind say they are ready

to improve the intelligence of robots by

imitating strategies that the human

brain uses in the decision-making

process

with this they plan to help a number of

recent discoveries in the field of

computational neurobiology

what will come out of this we will soon

learn take care of yourself subscribe to

the channel and give a thumbs up for the

video

let us know down in the comments what

did you find most impressive

 

You May Also Like

Avatar of admin

About the Author: admin