Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
The Singularity
A technically informed, but very speculative
critique of recent statements of e.g. Elon Musk,
Stephen Hawking and Bill Gates in the last year
“The Doomsday Invention”
http://www.newyorker.com/magazine/2015/11/23/d
oomsday-invention-artificial-intelligence-nick-
bostrom
CIS 391 - Into to AI2
Elon Musk Compares Building Artificial
Intelligence To “Summoning The Demon”
Musk: I think we should be very careful about artificial
intelligence. If I had to guess at what our biggest existential
threat is, it’s probably that. So we need to be very careful with
artificial intelligence.
I’m increasingly inclined to think that there should be some
regulatory oversight, maybe at the national and international
level, just to make sure that we don’t do something very
foolish. With artificial intelligence we’re summoning the
demon. You know those stories where there’s the guy with the
pentagram, and the holy water, and he’s like — Yeah, he’s sure
he can control the demon? Doesn’t work out.
MIT, October 2014 (from Techcrunch.com)
CIS 391 - Into to AI3
Elon Musk Compares Building Artificial
Intelligence To “Summoning The Demon”
Q: So I’ll take it there’ll be no HAL9000 going to
mars?
Musk: Heh. HAL 9000 would be easy [to deal with in
comparison to the AI he’s talking about]. It’s way
more complex… it’d put HAL9000 to shame. That’s
like a puppy dog.
CIS 391 - Into to AI4
Elon Musk: a Machine Tasked with Getting
Rid of Spam Could End Humanity
“I don’t think anyone realizes how quickly artificial
intelligence is advancing. Particularly if [the
machine is] involved in recursive self-improvement .
. . and its utility function is something that’s
detrimental to humanity, then it will have a very bad
effect,” said Musk.
“If its [function] is just something like getting rid of
e-mail spam and it determines the best way of
getting rid of spam is getting rid of humans . . . ”
Musk trailed off, as the crowd laughed.Vanity Fair New Establishment Summit October 2014
CIS 391 - Into to AI5
Stephen Hawking: AI could end human race
“The primitive forms of artificial intelligence we
already have have proved very useful. But I fear the
development of full artificial intelligence could spell
the end of the human race.
Once humans develop artificial intelligence, it would
take off on its own, and re-design itself at an ever
increasing rate.
Humans, who are limited by slow biological
evolution, couldn't compete and would be
superseded“
Stephen Hawking
CIS 391 - Into to AI6
Bill Gates
"I am in the camp that is concerned about super
intelligence. First the machines will do a lot of
jobs for us and not be super intelligent. That
should be positive if we manage it well. A few
decades after that though the intelligence is
strong enough to be a concern. I agree with Elon
Musk and some others on this and don't
understand why some people are not concerned.“
http://www.washingtonpost.com/blogs/the-switch/wp/2015/01/28/bill-gates-on-
dangers-of-artificial-intelligence-dont-understand-why-some-people-are-not-
concerned/
CIS 391 - Into to AI7
The Coming Technological Singularity: How to
Survive in the Post-Human Era: Vernor Vinge 1993
"Within 30 years, we will have the technological
means to create superhuman intelligence.
Shortly after, the human era will be ended."
CIS 391 - Into to AI8
Ray Kurzweil’s Singularity (Time Mag.)
CIS 391 - Into to AI9
The Singularity – Origin of the phrase
“His conversations with friends on scientific subjects
could last for hours. There never was a lack of
subjects, even when one departed from mathematical
topics.
…
One conversation centered on the ever accelerating
progress of technology and changes in the mode of
human, life, which gives the appearance of
approaching some essential singularity in the history
of the race beyond which human affairs, as we know
them, could not continue.”
– John von Neumann 1903-1957, Stanislaw Ulam
CIS 391 - Into to AI10
MIT Spinoff Funded $143M to Create
Sentient Computers
CIS 391 - Into to AI11
CIS 391 - Into to AI12
CIS 391 - Into to AI13
Deep Neural Nets – Breakthrough in
Speech Recognition Accuracy
CIS 391 - Into to AI14
Phone Error Rate
Training: 630 speakers x 10 sentences TIMIT
http://en.wikipedia.org/wiki/Deep_learning
Best HMM result
Best DNN result
Bye-bye
Moore’s Law..
CIS 391 - Into to AI15
Intel’s former chief architect Bob Colwell, head of DARPA’s
Microsystems Technology Office: “For planning horizons, I pick 2020 as
the earliest date we could call [Moore's law] dead. You could talk me
into 2022, but whether it will come at 7 or 5nm, it’s a big deal.”
The brain is complex in architecture..
CIS 391 - Into to AI16
Nancy Kanwisher, Functional specificity in the human brain: A window into the
functional architecture of the mind, Proceedings of the NAS, 2010
“Real soon now”
“By 1991, the impressive list of goals penned in
1981 for Japan's Fifth Generation Project had not
been met. Indeed, some of them, like "carry on a
casual conversation" had not been met by 2010. As
with other AI projects, expectations had run much
higher than what was actually possible.”
History of Artificial Intelligence, Wikipedia
Most of us accept that our lives are limited…. Not Ray
Kurzweil. Kurzweil, 65, decided decades ago
that mortality wasn't for him. He didn't have to die, and
he wasn't going to, if he could help it. Fortunately, he
believes he can help it—and he's been working
feverishly at the task of staying alive ever since.
“How long do you think you will live?” I asked Kurzweil
in a recent phone interview.
“I think I have a good chance—I would put it at 80
percent—of getting to the point where it becomes
indefinite, because you’ll be adding more time than is
going by to your remaining life expectancy.”http://www.slate.com/articles/technology/future_tense/2013/11/ray_kurzweil
_s_singularity_what_it_s_like_to_pursue_immortality.html
Then why believe this?
CIS 391 - Into to AI18
Three Bridges to Immortality
Bridge one: “ Our smartphones are
becoming extensions of our minds.”
Bridge two: “Kurzweil forecasts that
we’ll eventually be implanting computers and
nanobots in our bodies and brains to enhance
their natural functions.”
Bridge three: “And someday—by 2045, to be
precise—we’ll have machines so sophisticated
that we’ll essentially be able to back up our minds
to the cloud.”
CIS 391 - Into to AI19
A Less Radical AI Apocalypse
It could happen….
The worst case scenario for artificial intelligence run amok isn't killer
robots. It's widespread unemployment.
That's the view expressed by Andrew Ng, chief scientist at Baidu, a
Chinese-language Internet search engine….
Ng, a veteran of Google and Coursera, said he doesn't see a path from
the state of current technology to the sort of super-intelligence necessary
for malevolent cyborgs. But he does see reason to worry about the
economic impact of advances in computer learning and expert systems.
"If we succeed in building self-driving cars, that's 3.5 million truck
drivers who might have to find a new job," he said.
During the industrial revolution, Ng said, the population had 200 years
to shift from 98% farmers to 2% farmers. The transformation brought about
by computer technology "will be much faster.“
http://www.informationweek.com/it-life/artificial-intelligence-will-put-us-out-of-work/d/d-id/1318875
CIS 391 - Into to AI21
“Autonomous weapons: An open letter”
“Artificial Intelligence (AI) technology has reached a point
where the deployment of such systems is … feasible
within years, not decades, and the stakes are high:
autonomous weapons have been described as the third
revolution in warfare, after gunpowder and nuclear arms.”
“It will only be a matter of time until [autonomous
weapons] appear on the black market and in the hands of
terrorists, dictators wishing to better control their
populace, warlords wishing to perpetrate ethnic cleansing.
…” Stephen Hawking, Elon Musk, Steve Wozniak, Daniel Dennett, Noam Chomsky,
Martin Rees …
Stuart Russell, Nils Nilsson, Barbara Grosz, Tom Mitchell, Martha Pollack, Henry
Kautz,
Penn CIS: Aravind Joshi, Eric Eaton, <me>
CIS 391 - Into to AI22
It’s up to you…
CIS 391 - Into to AI23