and so it begins

[Suff]

Google has created a learning computer.

This is both good and bad. There is no talk about whether the computer is aware of anything beyond it's current environment, whether it has access to the local network, the Internet or anything else.

They say they want to create a computer with the mental age of a toddler. Unfortunately people often forget that toddlers have to learn all sorts of things. Like sight, walking, conforming, speaking etc and they spend about 50% of their day sleeping.

Computers, on the other hand, don't need sleep and can learn 24x7 at the speed of microprocessors. The more concerning thing is that already business is queuing up to get their hands on any tech which might ease the burden of programming and human knowledge.... See the slippery slope?

What happens when some geek who lives in a dark room decides to teach a program how to program itself on a range of systems, then gives it general instructions like "Go learn about the world", then lets it lose on the Internet which is the most powerful grid of computers ever seen.

Whilst there have been dozens of films and books about how we reach the age of "AI" and suddenly are destroyed by our creations, I think there is an inevitability to this. Scientist are too close to the trees to see the forest. It's how they work best.

Personally I believe that all self learning programs should have some core, unalterable, functionality which follows Asmiov's laws of robotics. Only in this way will we be protected from what we create.

1.A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2.A robot must obey the orders given it by human beings, except where such orders would conflict with the First Law.
3.A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

For Robot think "program", because a Robot is nothing more than a mobile and, perhaps, talking, program.

I shall be watching this closely from now on.

[TheOstrich]

You say this is both good and bad, Suff - what do you see as the upside in this development?

And how do you impose Isaac Asimov's laws on a programme which ultimately cannot be hack-proof?

Forgotten the title, but I'm reminded of that film where the kid hacked into the US Missile Command computer and played out war game scenarios with it ......

[Workingman]

Personally I believe that all self learning programs should have some core, unalterable, functionality which follows Asmiov's laws of robotics.

But what about those who disagree with your ideals, how do we deal with them? How do we control their programs?

[Suff]

Hanging?

Joking aside, there will come a time where creating these kind of programs is seen to be damaging to the entire human species. Our usual penalty for that is death.

[Aggers]

God help us when Science Fiction really becomes Science Fact.

We're halfway there already.

[Suff]

It doesn't have to be like that. All the teams involved in this have to do is think about the possible impacts first and then put in place checks and firewalls to stop it.

The problem is that nobody wants to fund that kind of work and teams on the cutting edge deem it to be limiting.

Responsible science is gone. It's been replaced with money and kudos.

[Workingman]

What about this? Brain control.

Now this is clever, and the possibilities extending from it are unquantifiable.

[Suff]

This kind of research has been expanding exponentially. This site is dedicated to uncovering abuse. Paranoid? Or not paranoid enough?

More about the science of it.

What we were seeing in the video was a very small part of the science being used. That entire cap was being used to track two movements, moving one of two balls up or down. It was, literally, that simple and the person controlling had to train themselves to consciously focus on the actions they would carry out if they were actually reaching out to move them.

interesting research though.