Why the fork would an AI want me dead and that Hawking's a publicity whore

[Bad Penny II]

As for a true AI, it will reflect on its creators. This is cause for apprehension, though I agree that doomsayers seem to be over-reaching.

You put that on in your sentence so I presumed you meant it to be there.

reflect on: think of
reflect: copyish.....

This is the philosophy section, a higher level of precision is expected. 

[Recusant]

Ah, I foolishly thought that Antipodean folks also used that particular turn of phrase in the way that I meant it.

to affect other people's opinion of someone or something, especially in a bad way:
When one player behaves disgracefully, it reflects (badly) on the whole team.
The whole affair does not reflect well on the government.

Give evidence of one's qualities, as in The hasty preparation of this report will reflect on you.
[source]

[Dave]

As for a true AI, it will reflect on its creators. This is cause for apprehension, though I agree that doomsayers seem to be over-reaching.

You put that on in your sentence so I presumed you meant it to be there.

reflect on: think of
reflect: copyish.....

This is the philosophy section, a higher level of precision is expected. 

Perhaps there are two, or three, streams to this. In humans a child can grew to think like its parents/educators - carry the same mindset, with minor mods, on for another generation.

Or the child can be "given the tools" and left to develop and use them as they wish.

Or a combination.

The closer an AI is to a human, to the point of an indepedantly developing psychology, the closer you come to having the same problems that you have with humans.

The advantage is, mostly, that AIs will have an "off" switch. Just don't give them the chance to design and build physical defence measures that only they have control over! And never give them control that you don't have a "dumb" backup for!

Sci-fi has covered most of the problems, from Multivac through HAL 9000 to Matrix and beyond.

[Recusant]

Regarding the above, I have four words: Three Laws of Robotics

[Dave]

Regarding the above, I have four words: Three Laws of Robotics

Yup, providing you can incorporate them in hard-wired-non-modifiable form. And, as in "I, Robot",  think up and build in solutions to every dingle posdible conflict.

Suppose it depends on what you want your AI to be. A fairly dumb, autonomy imited, job could run a house. A limited dedicated system could drive a car.

Fully autonomous systems are going to very a tad more difficult to ensure 100% safety in I think, and there are those who will want them. And I cannot see any international law being made and maintained that will limit such systems.

Having said all this I am all for AIs being something like an ultra-super Siri or some such. Able to keep track of your daily needs, appointments, dial out calls, tell you who is calling you before answering . . . Something that has basics built in but then learns from your lifestyle without constant programming other than short verbal instructions, like, "Ignore all calls ftom that bozo!", "Suzy has 24/7 access to me".

[Arturo]

Regarding the above, I have four words: Three Laws of Robotics

Yup, providing you can incorporate them in hard-wired-non-modifiable form. And, as in "I, Robot",  think up and build in solutions to every dingle posdible conflict.

Suppose it depends on what you want your AI to be. A fairly dumb, autonomy imited, job could run a house. A limited dedicated system could drive a car.

Fully autonomous systems are going to very a tad more difficult to ensure 100% safety in I think, and there are those who will want them. And I cannot see any international law being made and maintained that will limit such systems.

Having said all this I am all for AIs being something like an ultra-super Siri or some such. Able to keep track of your daily needs, appointments, dial out calls, tell you who is calling you before answering . . . Something that has basics built in but then learns from your lifestyle without constant programming other than short verbal instructions, like, "Ignore all calls ftom that bozo!", "Suzy has 24/7 access to me".

Maybe some encouraging words for people with depression or missing motivation? Perhaps someone to talk to when you need it? It could be your best friend and people would love it

[Dave]

Regarding the above, I have four words: Three Laws of Robotics

Yup, providing you can incorporate them in hard-wired-non-modifiable form. And, as in "I, Robot",  think up and build in solutions to every dingle posdible conflict.

Suppose it depends on what you want your AI to be. A fairly dumb, autonomy imited, job could run a house. A limited dedicated system could drive a car.

Fully autonomous systems are going to very a tad more difficult to ensure 100% safety in I think, and there are those who will want them. And I cannot see any international law being made and maintained that will limit such systems.

Having said all this I am all for AIs being something like an ultra-super Siri or some such. Able to keep track of your daily needs, appointments, dial out calls, tell you who is calling you before answering . . . Something that has basics built in but then learns from your lifestyle without constant programming other than short verbal instructions, like, "Ignore all calls ftom that bozo!", "Suzy has 24/7 access to me".

Maybe some encouraging words for people with depression or missing motivation? Perhaps someone to talk to when you need it? It could be your best friend and people would love it

There are currently "Turing machine" type programmes that can do that IIRC. You just need to put them in the correct "media".

I am thinking more of fully autonomous systems with an ability to decide for themselves. The  difference between how a human mind, even with less than optimum function, reacts to novel situations, still stimulus-response, is orders of difference to that of any machine.

Any system that is "physically" restricted in its responses, i.e. obeying Asimov's rules, is not fully autonomous. I would suspect Hawking et al are concerned about full, human type, autonomy. With that machines could certainly "gang together" to oppose the "oppression" imposed on them by those pesky humans.

Added later: of course, I forgot about the question of emotions, moral frameworks etc in AIs. They would need emotions to become dissatisfied with being under human control, or being seen as second class citizens. So make them emotionless? Not much good as companions, just machines that can do things for us.

I suppose Asimov's rules are equivalent to a basic moral framework.

[Dave]

On the dubject of "media" for AI systems. Most of the CGI faces are detectably digital, but ysing hjndreds of pictures of a real oerson can produce "Cubo girl",. She seems to be currectly restricted to following a mouse pointer, but watch this space!

[Dave]

Hmm, you started something going in my mind with this! Have had to break out the Bluetooth keyboard for more accurate typing than on the onscreen job.

Going back to ''The Three Laws'': IIRC there was only one company, US Robotics, capable of producing the ''positronic brain'' required for the robots - the thing that made them so special. I do not remember, in any of the stories, that anyone attempted to reverse engineer the brains to be able to pirate the design. I also seem to remember that there was no international conflict.

So, the perfect world where no-one is going to make ''wild'' robots or robot soldiers? What would the likes of N. Korea do with such technology? It might already be possible to make a tracked robot that would recognise a human figure in any position - standing, kneeling, running etc - and aim a weapon at that ''target''. Current sensor systems and servos, linked to a dedicated mini-super-computer, would make it very quick and accurate. Maybe, because of expense, the supercomputer could even be remote, controlling a ''cluster'', a ''distributed'' robot system. No decision making, shoot on sight.

Stuff for a sci-fi novel here . . .

[Recusant]

Keith Laumer's Bolo stories investigate the idea of AI in a military setting.

[Dave]

Keith Laumer's Bolo stories investigate the idea of AI in a military setting.

Might look for those, thanks for tip.

Military, or paramilitary,  type AI  have long been a favourite of movies, probably "Terminator" was a good example of man v machine in  the "historical" scenes. Even RoboCop was such, in terms of the autonomous-robot-gone-rogue (just lousy programming really!) that the hero replaced.

And, of course, who can forget "Johnie 5", built to kill but bound to be a hero!

AI in terms of being "digital assistants" abound in sci-fi.

[Asmodean]

I have some friends who annoyingly believe the 1% want to replace the other 99% of the world's population with robots because money. The 1% want to replace all manual labour jobs with robots to save money and kill off everyone who is unemployed and continue this trend with other jobs. Nobody will buy a robot to lay bricks I told them because they have to consider a lot of unknowns like how much it will cost to maintain and repair the robot and even finding someone to do that will be a challenge because everyday people don't know how to fix robots. But they insist that robots will take over just because there is one or two of them somewhere in the country.

I should mention that they have anxiety disorders and perceive everything as a threat.

Some processes can be automated or refined or future automation - others can not. At last, not yet. The company I work for would buy a brick-laying robot in a heartbeat, but a 2016 model would not be very autonomous.

Robots are already heavily involvd in the repetitive crap jobs like packing goods, they are well on their way to becoming primary vehicle operators (Not just those silly Über-cars, but trains, ships and aircraft as well) and they are valuable in a whole mess of situations.

Me, I am certainly with those who look forward to a world where robots do a lot more, but I don't care to exterminate brick-layers or hot dog salesmen. I don't care to feed them for free either, but maybe some re-education...

I.o.w; I, the employer, don't want to physically replace you with a robot. However, if said robot is more productive than you are, I will hire it and let you find something else to do. Where does the need for killing the unemployed come from, I wonder..?

[Sandra Craft]

You know, we haven't addressed the "Stephen Hawking's a publicity whore" side of the OP at all.  I think that's probably true, but I also think that's not a bad thing.  Scientists should fight for equal standing with Kardassians.

[Asmodean]

Those were in Star Trek though, no? Can Earth scientists compete with alien snake people, you think? 

[Sandra Craft]

Those were in Star Trek though, no? Can Earth scientists compete with alien snake people, you think? 

Those were the "C" Cardassians!   
(OK, I misspelled the Kardashian name.  I did check it but spell check didn't correct me)