YouSub
Why You Shouldn’t Fear Artificial Intelligence

en English

Why You Shouldn’t Fear Artificial Intelligence English subtitles

0 downloads · 47 lines · added Feb 14, 2016

Report a problem
What's wrong?
Advertisement

English transcript

A.I. might be a human-destroying killer, but without ANY A.I., the world wouldn't be a
very awesome place.
Stephen Hawking famously said to the BBC, "The development of full artificial intelligence
could spell the end of the human race[1]." Elon Musk has said it's like, "summoning the
demon[2]." And yet, earlier this month, Elon Musk, Stephen Hawking and important minds
from around the world signed an open letter from the Future of Life Institute encouraging
a future for the technology. The letter was widely interpreted by the media[3],[4] to
indicate these super-smart peeps are saying A.I. is going to KILL US ALL, but that's not
entirely accurate.
It says, right in the open letter[5], "The potential benefits are huge, since everything
that civilization has to offer is a product of human intelligence," what they ARE saying
is how A.I. may change the way we live, work, and affect the planet. A.I. could solve problems
in ways we can't conceive, and if done right might create a better world for us in the
process! Might, being the operative word here…
At the moment, A.I. isn't dangerous in itself unless you're living in the realms of films
and television. The open letter is some of the greatest humans NOW wanting to make sure
the A.I. of the FUTURE will be programmed to concern human safety. Musk donated 10 million
dollars alone to safety of A.I.[6] because he knows: computers are dumb. We have to teach
them morality, ethics and empathy, just like we train it to drive, speak, move cargo or
play piano. We have to recognize, the humans PROGRAMMING the A.I. have biases. If we create
a banking A.I. using current lending data as data points, it might lend more to white
people than black people; because of the racism of the people who "taught" the computer! It's
all in how we program it[7].
This open letter acknowledges[8] that A.I. is beneficial, and it it's a pledge to make
sure that's a priority in the future too. Early A.I. work was mainly done by computer
scientists, but it's now interdisciplinary; encompassing biology, medicine, law and ethics,
psychology and so on. Researchers are saying A.I. should make the world safer and pledging
to that fact.
As of now, the most advanced artificial intelligences are still pretty basic, working in areas like
language processing, photo and face recognition, shipping and logistics, or in driverless cars.
To say these are "intelligent" isn't entirely correct. Though often A.I. do have some machine
learning, computers only know what we teach them. For example, the brain of a driverless
car can take you to the store, but it can't shop for groceries; the variables are completely
different.
As of 2013[9], the most advanced A.I. in the world, according to Business Insider, was
ConceptNet4. It was built by M.I.T. and had the intelligence of a preschooler, it could
explore a bit on it's own, but mostly needed humans to input information for it to learn.
They're now up to ConceptNet5.3, and it's a free download[10]! Sure, a full A.I. might
be bad, but these rudimentary A.I. are nothing but helpful and should continue to be! Stephen
Hawking himself uses a computer with type of A.I. to speak[11] called Assistive Context
Aware Toolkit (ACAT[12]). It uses his previous writings and communication to predict what
he's going to want to say. That doesn't sound too dangerous… and without the A.I. being
able to learn, it wouldn't be much help. Hawking couldn't even TELL us to watch out for A.I.
without A.I.!!
Hawking, Musk, and researchers from the largest and smartest technology firms and universities
in the world aren't saying, "Run for your lives!" They're saying "Run the right algorithms!"
What do you think about Artificial Intelligence? Would you want it?

Need subtitles from another video?