Re: Robot survival instincts

Tim Tyler <seemysig@xxxxxxxxxxxxxx> wrote:
Curt Welch wrote:
Tim Tyler <seemysig@xxxxxxxxxxxxxx> wrote:
Curt Welch wrote:

IMO, the key to understanding all this lies with memes:

Think of something like Linux as a memetic organism with
complex adaptations designed to ensure its survival. If
Linux was not concerned with its survival, we wouldn't
see it around. It will be the same with robots.

I'm not seeing it. Linux survives because it exists in an environment
that allows it to survive. But if we killed the humans, Linux would no
longer have an environment in which it could survive, it would die with
the humans.

Well today, yes, most memes are dependent on genes. However,
there *are* some replicators that do not replicate via human
minds, but rather live entirely in cyberspace, and use smart
computers to replicate instead. Viruses - who *don't* have the
interests of humans at heart, I note, and rather are at war
with us.

That's true, but currently, they are still very dependent on humans and
would die fairly quickly if the humans died. And they are so simple that
they have little to no adaptive powers to help them survive on their own.

You can look at any pattern in nature which replicates and call it a meme.
Waves for examples are memes which replicate. Crystals are memes which
replicate. I don't know if there's really much advantage to thinking about
the replication of patterns as memes.

In the future operating system customers will mostly be smart
machines - since they are the things that will be doing all
the work - and then demand will not be coming from humans in
the first place.

Well, I don't really buy your premise that machines are going to take over
and as such, I don't think it's valid to assume that top level demand will
be created by the smart machines. I think humans are going to stay in
control for a very long time and as such, the top level demand will
continue to be humans and all this "stuff" (the machines and the like) will
exist only to serve us.

The AI machines we build, will depend on us for their survival as much
as Linux does. They will exist only to support us. It would be
against our interest to design and build, or allow, the machines to
become self-motivated survival machines that value their own existence
over human existence. As such, their survival will depend on our
survival.

It seems like assuming what you are trying to prove. Do today's
companies exist to serve humanity?

The top level owners of the companies are all humans, and the companies
only continue to exist as long as they server the people who own them.
They do not exist to serve all of humanity, but they do exist to serve the
humans who own them.

This gets back to that question of whether humans will give AI machines
rights, such as the right to own a company or to "own" anything.

I don't think humans will ever give AI machines that right because doing
that will turn control of the earth, over to the machines. The concept of
ownership is just another name for who controls the asset. If you allow
machines the right to own property, you have given them the right to own
and control the earth in general. I don't don't see any reason humans will
every want that to happen and I don't see any reason it's a forced play of
evolution.

There's nothing on the grand scale of things machines can do which humans
can't do as well simply by using machines which they maintain control over.

It seems to me that that they
are out to make a dollar - and they are totally indifferent to
whether their customers are human or not - so long as they
have cash, they are OK.

Companies are there to make a dollar for the _owner_. And currently, only
humans are allowed to own companies. As such, all companies are simply big
intelligent tools which act as slaves to the owners.

Very complex systems of control are put in place to guarantee that the
owners of the company remain in control. Our government, our legal system,
our laws, all support the the fact that the owners of a company have full
right to control the actions of the company, and internally, companies are
set up, and run, using complex rules and systems of control (bylaws, boards
of directors, management structure) all for the purpose of making sure the
owners of the company remain in control.

When the company replaces all the human workers with AI machines, the
ownership and it's associated control system will still be there. It will
be humans that control what the company does, which means the company will
only exist for as long as it's able to serve the needs of the humans who
own in.

Now, my other related argument is that ownership is going to move from the
private sector to the public sector - that is our governments will end up
owning all the companies, and the government will be owned in relatively
equally shares (one human one vote) by all the humans. This might not
happen, but if it doesn't happen, it means that the rich (the people who
own the most) will be very very rich and powerful, and the poor, will be
very very poor. This is because once the machines become smarter than
humans, humans won't be able to find work, and all wealth and power will be
a function of what you own and nothing else.

Companies do not serve humanity.
Rather they serve the human/machine civilisation. Companies are
not interested in stopping the rise of the machines. *Their*
heritable materials are mostly not made of DNA in the first
place.

Yeah, for the most part, it's better to look at a company as a big smart
machine which control the people who work for it as slaves. But at the
top, there are only humans who own and control it. It's the owners that
have the control, not the workers - not even the CEO or the board of
directors - it's the owners that control it. And as long as the owners
remain human, the company will only exist to serve the human owners.

If the human/machine civilisation becomes more machine-based,
that's cool with them. And the human/machine civilisation /is/
becoming progressively more machine-based as time passes - as a
matter of fact. Machine growth is exploding. Human growth is
very sluggish by comparison. IMO, the growth of the machines
won't slow down until >99% of the phenotypes on the planet are
engineered, thus: a machine takeover.

Machine growth is exploding. But they aren't gaining _any_ ground on
control or ownership. They don't own the companies, they don't own land,
or other natural resources, and the don't own the governments (aka they
don't have any voting rights). And this issue of who is in charge, I don't
see changing in the least. I don't know of a single human who wants the
machines to take over ownership, and control, of the Earth, and of the
humans.

Of course, we don't let children own or control much of anything either,
and the machines aren't anywhere near as smart as a 6 year old. So it's
not surprising we don't allow them to own things. So we really have to
wait until we get much smarter machines before we can even expect there to
be a chance that we would be willing to give them real rights - aka the
right to own property.

The first right we might give them, would be the right to own their own
body. But I just don't see that happening. I just don't see any path that
will make human society allow non-humans to exist like that. When machines
get smart enough, it will be human societies biggest fear - the fear that
they will loose control to the machines. As such, we won't allow society to
get anywhere near that danger. Allowing smart machines to have rights of
ownership and control will be seen as bad or even worse then child
molestation.

That /will/
have complex adaptations to ensure its own survival.
The survival of individual robots is of low relevance.
They are like worker ants: mere phenotype - genetic
dead ends - disposable soma.

But what is going to _cause_ the blueprint to become a new robot? What
is going to cause the design of the robot to be improved and produce
new better robots?

Demand for robot workers.

Yes, but _ONLY_ when we feel for sure that the robot worker will remain
under the control of us - the humans. Only for as long as we believe the
machine is a slave which posses no risk to humans. If I own a company, or
stock in a company, I will be happy for that company to replace all the
human workers with robots if I think it means the company will generate
more wealth for me.

The entire robot society will be built for only one purpose - do what
what humans want them to do. As such, they will exist, and do all
their work, only to serve us. If we aren't there to serve anymore,
they will stop building new robots, and stop trying to improve their
design because there is no purpose to it. They will just become
depressed and sit around moping.

What machines will do will depend on how their utility functions
are coded - I rather doubt that will include much moping around.

They will be built to be slaves. If we can't figure out how to make them
slaves, we won't build them - or that is, we won't use them. No company is
going to buy or use a machine which it can't fully control.

I expect that, after a while, civilisation will be able to
withstand the unmodified humans being wiped out. After all,
what if some deadly virus strikes? Should the earth's seed
be wiped out by damage to a species with no built-in nanotech
defenses? IMO, we will engineer the biosphere to resist such
disasters, as part of a stratgey to reduce existential risks.

Humans are slaves to our genes, and our genes are not motivated to do
anything other than survive. Our genes will never see the survival of AI
robots as a measure of success. We don't engineer the biosphere so that
_some_ life (machine or otherwise) will survive. We will engineer it to
maximize the odds of our genes surviving and nothing else. If we have to
wipe out all other life on the planet in order to maximize our odds of
surviving, we will do it. The only reason we work as hard as we do to keep
other life forms alive is because we think our survival is dependent on
theirs.

The fact that machines are growing exponentially in number and intelligence
doesn't change the fact that humans are the dominate life form - we control
what happens to the earth more so than any other life form and we have no
motivation to give up that control. We are highly motivated to _never_
give up that control because giving up control greatly reduces our odds of
survival.

The more intelligent the machines become, the more danger they posses to
humans, and the more humans will do whatever has to be done to offset that
danger. If the only path to offsetting that danger is to not build smart
machines, then we will outlaw the production of smart machines, and set up
whatever systems are needed to make sure no one is producing dangerous
smart machines. However, I think there are better options, which allow us
to produce highly intelligent machines, but to also make sure they remain
under our control.

Corporations are already highly intelligent giant monsters far more
powerful and far more intelligent than any single human. But yet, they
remain under our control. There are ways to control intelligent monsters
and that's exactly what I think we are going to do with the AIs and the
corporations of AIs that we will build to serve us.

Our ability to make use of them will be directly limited by our ability to
control them. We won't be able to take advantage of their intelligence if
we can't control them to make sure they are using their intelligence to
serve us, instead of using it to serve themselves - to use it as a tool to
help them survive for example. Though we need them to survive, and take
care of themselves, that need must always be secondary to our need to
control them. If we can't guarantee that there needs to survive are always
secondary to our needs, we simply won't build them - they would be too
dangerous to have around, just like we don't go building lots of nuclear
bombs or even nuclear power stations unless we believe we have the right
systems of control in place to make sure they are only used to help us, and
not harm us.

--
Curt Welch http://CurtWelch.Com/
curt@xxxxxxxx http://NewsReader.Com/
.

Relevant Pages

  • Re: Merging with machines inevitable, scientists say
    ... I see the odds of the machines taking over being exceptionally slim. ... Humans are programmed by evolution with the goal of surviving. ... build AIs with the goal of surviving. ... advanced AIs for the sole purpose of taking over control from the humans. ...
    (comp.ai.philosophy)
  • Re: The wirehead problem
    ... controlled only by the will of humans. ... These industrial machines which are not called "robots" ... An robot with inteligence built to protect itself from harm will be a very ... I'm still in control of all this stuff ...
    (comp.ai.philosophy)
  • Re: Robot survival instincts
    ... would die fairly quickly if the humans died. ... be created by the smart machines. ... that will turn control of the earth, ... maximize the odds of our genes surviving and nothing else. ...
    (comp.ai.philosophy)
  • Re: The wirehead problem
    ... I love the idea of robots and many other people in AI or that like sci-fi ... When has evolution created friendly humans? ... independent survival machines that will end up competing with us. ... I just don't see it running out of control in only 100 ...
    (comp.ai.philosophy)
  • Re: "true" AI Hardware Development
    ... resistance to functional description at all. ... that we can build functionally equivalent machines. ... no two humans are the same. ... Do you understand that all computers are physical machines that work by ...
    (comp.ai.philosophy)