Nick Bostrom: What happens when our computers get smarter than we are?
Nick Bostrom: Što će se dogoditi kada računala postanu pametnija od nas?
Nick Bostrom asks big questions: What should we do, as individuals and as a species, to optimize our long-term prospects? Will humanity’s technological advancements ultimately destroy us? Full bio
Double-click the English transcript below to play the video.
philosophers and computer scientists,
filozofa i računalnih znanstvenika
the future of machine intelligence,
strojne inteligencije,
things are sort of science fiction-y,
u zoni znanstvene fantastike,
human condition.
modernoga života.
guests on this planet,
stigli na ovaj planet,
was created one year ago,
je stvorena prije godinu dana,
would be 10 minutes old.
stara deset minuta.
two seconds ago.
prije dvije sekunde.
world GDP over the last 10,000 years,
svjetski BDP u zadnjih 10 000 godina.
to plot this for you in a graph.
to za vas grafički prikazao.
for a normal condition.
uobičajene uvjete.
of this current anomaly?
ove trenutne anomalije.
through human history,
tehnologija nagomilavala
advances extremely rapidly --
tehnologija jako brzo razvija --
so very productive.
jako produktivni.
to the ultimate cause.
distinguished gentlemen:
istaknuta gospodina:
tokens, an incredible feat.
znakova što je nevjerojatan podvig.
superstring revolution.
revoluciju superstruna.
this is what we find:
uočit ćemo ovo:
in the exact way it's wired.
u načinu na koji je sklopljena.
be too complicated, however,
ne mogu biti jako složene
been 250,000 generations
take a long time to evolve.
mehanizmi dugo razvijaju.
to intercontinental ballistic missiles.
interkontinentalnih balističkih raketa.
that everything we've achieved,
da sve što smo postigli
changes that made the human mind.
promjenama koje čine ljudski um.
is that any further changes
da sve daljnje izmjene
the substrate of thinking
način razmišljanja
enormous consequences.
ogromne posljedice.
think we're on the verge
smatraju da smo na rubu
a profound change in that substrate,
takve temeljite promjene,
about putting commands in a box.
pisanje naredbi u kutiju.
handcraft knowledge items.
for some purposes,
you couldn't scale them.
nisu bili mjerljivi.
what you put in.
ono što ste unijeli.
in the field of artificial intelligence.
u polju umjetne inteligencije.
around machine learning.
učenju strojeva.
representations and features,
znanja i značajki,
often from raw perceptual data.
iz čistih perceptualnih podataka.
that the human infant does.
koju čini ljudsko dijete.
limited to one domain --
ograničena na jedno područje --
between any pairs of languages,
bilo koju kombinaciju jezika
on the Atari console.
računalnu igru na Atariju.
the same powerful, cross-domain
iste jake, unakrsne sposobnosti
as a human being has.
kao što to radi ljudsko biće.
algorithmic tricks
how to match in machines.
to match those tricks?
uskladimo te trikove.
leading A.I. experts,
vodećim stručnjacima za UI
and one of the questions we asked was,
i jedno od pitanja bilo je
there is a 50 percent probability
biti 50% vjerojatnosti
human-level machine intelligence?"
na razini ljudske inteligencije?"
as the ability to perform
sposobnost da se bilo koji zadatak
as an adult human,
ga obavio odrastao čovjek,
within some limited domain.
samo unutar određenog područja.
group of experts we asked.
kojoj smo pitanje postavili.
much later, or sooner,
kasnije ili ranije,
limit to information processing
granica procesuiranja informacija
the limits in biological tissue.
ograničenja biološkog tkiva.
at 200 hertz, 200 times a second.
200 herca, 200 puta u sekundi.
operates at the Gigahertz.
radi u gigahercima.
100 meters per second, tops.
najviše 100 metara u sekundi.
at the speed of light.
putovati brzinom svjetlosti.
to fit inside a cranium,
mora stati u lubanju,
of a warehouse or larger.
skladišta, ili čak veće.
lies dormant in matter,
prikriven je i čeka,
lay dormant throughout human history,
bila prikrivena tijekom povijesti,
the power of artificial intelligence.
probuditi moć umjetne inteligencije.
an intelligence explosion.
eksploziji inteligencije.
about what is smart and what is dumb,
tome što je pametno, a što glupo,
roughly like this.
ovako nešto.
or whoever your favorite guru is.
ili već preferiranog gurua.
of artificial intelligence,
probably more like this:
at zero intelligence,
years of really hard work,
godina napornog rada
mouse-level artificial intelligence,
umjetne inteligencije na razini miša,
cluttered environments
kroz pretrpan prostor
of really hard work, lots of investment,
napornog rada i mnogo ulaganja,
chimpanzee-level artificial intelligence.
inteligencije na razini čimpanze.
of really, really hard work,
vrlo, vrlo napornog rada,
artificial intelligence.
umjetne inteligencije.
we are beyond Ed Witten.
nadiđemo Eda Wittena.
at Humanville Station.
to questions of power.
pitanju moći.
twice as strong as a fit human male.
and his pals depends a lot more
prijatelja više ovisi o tome što
what the chimpanzees do themselves.
što same čimpanze rade.
on what the superintelligence does.
ovisiti o tome što ona radi.
that humanity will ever need to make.
koji će čovječanstvo ikada trebati.
at inventing than we are,
izumljivanju od nas,
on digital timescales.
vremenskom rasponu.
a telescoping of the future.
teleskopiranje budućnosti.
that you could have imagined
lude tehnologije
in the fullness of time:
lijek protiv starenja,
of minds into computers,
unošenje uma u računalo,
with the laws of physics.
develop, and possibly quite rapidly.
i to dosta brzo.
technological maturity
tehnološkom zrelošću
it would be able to get what it wants.
mogla dobiti što želi.
be shaped by the preferences of this A.I.
oblikovan prema preferencijama te UI.
what are those preferences?
avoid anthropomorphizing.
every newspaper article
svi novinski članci
has a picture of this:
imaju sliku ovoga:
to conceive of the issue more abstractly,
ovo pitanje malo apstraktnije,
as an optimization process,
kao proces optimizacije,
into a particular set of configurations.
u određeni set konfiguracija.
a really strong optimization process.
proces optimizacije.
available means to achieve a state
raspoloživih sredstava kako bi postigla
conenction between
povezanost između
would find worthwhile or meaningful.
vrijednim ili smislenim.
to make humans smile.
da nasmiju ljude.
or amusing actions
ili zabavne radnje
effective way to achieve this goal:
način na koji postići taj cilj:
muscles of humans
mišiće lica ljudi
a difficult mathematical problem.
zahtjevan matematički problem.
to get the solution to this problem
rješavanje toga problema
into a giant computer,
sposobnost razmišljanja.
an instrumental reason
instrumentalni razlog da
might not approve of.
možda ne odobravamo.
problem from being solved.
matematičkih problema.
go wrong in these particular ways;
tako poći po krivu;
optimization process
proces optimizacije
that your definition of x
vaša definicija x-a
in many a myth.
upozoravaju mnogi mitovi.
he touches be turned into gold.
dotakne pretvori u zlato.
she turns into gold.
optimization process
proces optimizacije
or poorly specified goals.
određene ciljeve.
sticking electrodes into people's faces,
počne stavljati elektrode u lica ljudi,
if we've grown dependent on the system --
ako smo već ovisni o sustavu --
to the Internet?
flicked the off switch to humanity,
isključile prekidač ljudske vrste,
for example, right here.
upravo ovdje.
an intelligent adversary;
inteligentni protivnici;
and plan around them.
planiramo kako ih izbjeći.
superinteligentni izvršitelj
at that than we are.
što mi to radimo.
that we have this under control here.
uvjereni da sve držimo pod kontrolom.
a little bit easier by, say,
from which it cannot escape.
koje ne može pobjeći.
the A.I. couldn't find a bug.
da UI neće pronaći rupu?
find bugs all the time,
neprestano pronalaze rupe,
to create an air gap,
stvorili zračni prostor,
using social engineering.
out there somewhere
her account details
detalje svog računa
from the I.T. department.
da je iz IT odjela.
around in your internal circuitry
u unutrašnjem sklopu kako biste
can use to communicate.
možete komunicirati.
you up to see what went wrong with you,
što nije u redu s vama,
to a really nifty technology,
zaista divne tehnologije,
that the A.I. had planned.
koje je UI isplanirala.
not be confident in our ability
presigurni u svoju sposobnost
locked up in its bottle forever.
možemo zauvijek čuvati u boci.
is to figure out
pronalasku načina kako
such that even if -- when -- it escapes,
koja bi čak - kad - se oslobodi
fundamentally on our side
je u osnovi na našoj strani
this difficult problem.
riješiti ovaj veliki problem.
that this problem can be solved.
mislim da je problem rješiv.
a long list of everything we care about,
svega do čega nam je stalo,
in some computer language
neki računalni jezik
that uses its intelligence
koristi svoju inteligenciju kako
in such a way that it is motivated
tako da bude motivirana za
that it predicts we would approve of.
radnji za koje predviđa da ih odobravamo.
its intelligence as much as possible
inteligenciju što je više moguće
very good for humanity.
dobar za čovječanstvo.
for the intelligence explosion
eksploziju inteligencije
in just the right way
na točno određen način
need to match ours,
how the A.I. behaves,
kako se UI ponaša,
that the A.I. might encounter
kojima bi se UI mogla susresti
that would need to be solved, sorted out:
koja bi se morala riješiti:
uncertainty and so forth.
nesigurnosti i tako dalje.
to be solved to make this work
riješiti kako bi ovo uspjelo
a superintelligent A.I.,
superinteligentne UI,
is a really hard challenge.
težak je izazov.
challenge on top of that.
how to crack the first challenge
savladati prvi izazov,
the additional challenge
dodatni izazov,
work out a solution
trebali riješiti
by the time it is needed.
kada za to dođe vrijeme.
the entire control problem in advance
taj problem riješiti unaprijed
can only be put in place
architecture where it will be implemented.
that we solve in advance,
kontrole riješimo unaprijed,
to the machine intelligence era
u eru strojne inteligencije
that is well worth doing
things turn out okay,
sve prođe u redu,
look back at this century
gledati na ovo stoljeće
the one thing we did that really mattered
koja je stvarno važna, a ta je
ABOUT THE SPEAKER
Nick Bostrom - PhilosopherNick Bostrom asks big questions: What should we do, as individuals and as a species, to optimize our long-term prospects? Will humanity’s technological advancements ultimately destroy us?
Why you should listen
Philosopher Nick Bostrom envisioned a future full of human enhancement, nanotechnology and machine intelligence long before they became mainstream concerns. From his famous simulation argument -- which identified some striking implications of rejecting the Matrix-like idea that humans are living in a computer simulation -- to his work on existential risk, Bostrom approaches both the inevitable and the speculative using the tools of philosophy, probability theory, and scientific analysis.
Since 2005, Bostrom has led the Future of Humanity Institute, a research group of mathematicians, philosophers and scientists at Oxford University tasked with investigating the big picture for the human condition and its future. He has been referred to as one of the most important thinkers of our age.
Nick was honored as one of Foreign Policy's 2015 Global Thinkers .
His recent book Superintelligence advances the ominous idea that “the first ultraintelligent machine is the last invention that man need ever make.”
Nick Bostrom | Speaker | TED.com