[Ultron]

http://www.theguardian.com/technology/2015/feb/20/internet-of-things-artificially-intelligent-stephen-hawking-spike-jonze

Excellent topic worthy of a discussion. So, what will hapen?

And one note...

AI is broadly described as the ability of computer systems to ape or mimic human intelligent behavior.

This is why I think everyone has a problem...

[Art]

Fine. Here is my take (for the moment).

I believe that quote was incorrect as stated.

While that definition might apply to a chatbot or phone based app or digital assistant, the more hopeful and practical definition should read,"AI is the ability of a computer system to become intelligent, capable, self-correcting and self aware."

In other words, it WILL become intelligent and there is no mimicry involved...it simply will be intelligent through thought, emotional appropriateness and general behavior. It will be able to engage an average conversation or an advanced on based on theory. Any game in an attempt to prove it's computing power would be moot as it could easily beat the top opponent on any board game.

It will be capable of making decisions based on a special weighted system, always putting human life ahead of any other life, then stepping down the chain based on necessity or urgency, in case of disaster or danger. It would understand the need for self-sacrifice if no other recourse was available and all human weights had been made or exhausted.

An AI must think, be creative, compassionate, honest - for most occasions (except where life might be in danger), strive to become better and to survive.

(funny but for decades, people have wanted these things to be endowed with an AI and now that it's potentially within our grasp, they suddenly get scared that their wish might come true! Come on you can't have it both ways!) - Put up or shut up!

just a rant...there...I feel better now.

[Ultron]

I believe the media has too much power. People don't have their own opinion and thus it takes only 3 popular figures with the word 'science' in their biography to make everyone afraid (of anything they wish).

Back on topic, I believe the potential for new knowledge and good in this situation outweighs the potential of danger.

On the more specific topic of intelligent 'internet of things', I believe that would be the same as a robot, except scaled up (imagine the world and an android - android planet / android ecosystem? let's coin a term! ). So, according to some logic I guess it would be able to learn and evolve faster and better. Except the catch is somewhat large - it would have unprecedented power and control. Also all that data might be overwhelming and I doubt that even the largest computer clusters today could handle the amount of data.

[Don Patrick]

I read an article in the newspaper yesterday about people who have "smart thermostats" installed in their homes: All activity is passed to the power company, and one of the journalists tapped into that system simply by entering the name and address of his neighbour, so knowing when his neighbour was out by the fact that the heat was turned down. 75% of people refused to have free smart thermostats installed on the grounds that they wouldn't know who was going to do what with their private data.
These systems are great for burglars, and their security clearly hasn't been a priority yet. Now consider that controlling these systems would also be the greatest sport ever invented for hackers. Can we worry about that before we worry about A.I. with irrational grudges towards mankind?

[Ultron]

Excellent example Don. Yes, the danger from an A.I. controlling all of this is the same as that of a hacker - in both cases an entity gains massive control over some (to be in the future) key system. And when you think about it, there is nothing that can't be hacked - in time. So I guess the best security with smart homes is to have everything off the grid, so at least you won't have to worry about anyone hacking it wirelessly.

[Art]

I have a smart thermostat in my home that I installed and set up. It connects to my secure router and is accessible from anywhere with my smart phone, Provided I know (and I do) my password and access codes.
I can also turn some lights on and off or tell whether they are on or off from the phone.

I had a lot of things connected at one time but I've since lessened the numbers. I still like some control over some items, for now.

One of the easiest, cost effective methods for limited home automation is to but some lamp / appliance timers and some motion detector lights and plug-in units. Outdoor flood lights can come on with either a dusk to dawn timer or motion sensor as can backyard security lighting! Most of these are inexpensive and many can be had at your local "box" stores like Home Depot, Lowes, Wal-mart, Costco, BJ's, Sams Club, etc. Don't know what stores you might have of a similar nature where you live but check your local hardware or gadget stores or even online via eBay, Amazon or such.

[Ultron]

Art - if it takes a teenager to hack a secured Wi-Fi connection (back-track linux, anyone?), secured by WPA2-PSK - then it must take no more then a monkey to hack your thermostat.

I would love to have control over my house from anywhere in the world - but wireless access has it's price - and it's not a bargain.

[Art]

Gosh, then think of the damage that could be done by fooling with my thermostat! I'm floored!

First, they need to bypass several "doors", hardware and software and long passwords just to even find out that I have one much less screw with it.

Secondly, I really don't care. If they have nothing else to do with their time then they should really get a life and try some thermostats worth manipulating, like the freezers at local super markets all set to 65 degrees! Over several unnoticed hours that should cause the store some real damage.

So with me, I limit the connected things in my house to safe ones. I do not have those Door Locks that can be remotely unlocked / locked nor garage doors that can be raised / lowered via Internet command. While some do, I choose not to do so.

Thanks for your input! Any monkeys out there want to play with my thermostat? it goes from 58 to 74º.

[Don Patrick]

I have to agree along that line of thinking, it would be terribly difficult even for a superAI to kill someone by changing the thermostat to its lowest setting in winter or to its highest setting in summer. Thermostats are made for survivable temperatures only, so that's a decent failsafe for robot apocalypses and assassin hackers.

[ranch vermin]

When Don mention that thermostat thing, I got a little embarressed, because I was imagining some huge embarressing flaw in my product, I didnt notice.
On security,  that is definitely a big subject to think about before you release your item.  (Just think how much the dudes that made malwarebytes had to think about how hackers could hack their product off the computer.)   Its very good someone mentioned it, because its damn well needed to be thought about before you go release your ai out there,  if its something like heavy farm equipment or something dangerous like that,  better damn well have a hard think, even if its looking like its working.   Or if its some security system,  what if people can confuse the system somehow, or something like that,  especially ai security.   Or take advantage of its statistics database, to find how to rob a place!  use the security system against the security.

And on whats 'real artificial intelligence -  and whats fake artificial intelligence?'   the fact is its A - i.    dont forget what A stands for.  ARTIFICIAL.  how about going one step further, forget artificial, and why dont you get realistic,   its ALGORYTHMIC INTELLIGENCE.   so mimicry is actually the best tactic you can have unless your just pipe dreaming!

[Ultron]

I am not familiar with your security precautions Art nor what you have digitalized however my point was that regardless of whether it's a thermostat, an electronic door lock or security lasers in your backyard - it's hackable. No matter how advanced and heavy your security protocols are just remember that a man invented them, and the same man can break those codes in seconds.

Also, I believe it's possible to create a general decryption algorithm, something A.I. could use as a guide and way of thinking in order to test or at least consider all of the possible types of encryption methods. So, the future would not surprise me with such intelligence.

[Don Patrick]

Just think how much the dudes that made malwarebytes had to think about how hackers could hack their product off the computer.

Coincidentally, my parents currently have a virus on their computer that is blocking malwarebytes from activating or updating.

[Art]

Oh I absolutely agree Ultron. I never said it wasn't hackable, I just mentioned, what's the point? There's nothing to be gained except for a lot of frustration for the average kid hacker trying to gain access to my home thermostat. I can also easily manually override any erroneous commands.

In fact, your door lock on your house or apartment is hardly secure! Anyone with basic locksmith knowledge can be inside in a matter of several seconds or at least half a minute if he/she is slow.

Security is just an illusion and it is peddled to anyone who wishes to believe that it exists (much like a lot of things).

The idea behind hacking usually revolves around the adage, "Where there is risk, there is reward!" or Nothing ventured - nothing gained!

Instead of proclaiming the word Security I'd prefer to use the word, Deterrent and that can really cause one to wonder.

[Ultron]

And I also agree that security is a mere illusion so it is worth noting here that you shouldn't invest in the material world, yet you should invest in knowledge since it is the only thing that can't be stolen from you... Yet.

But one should also realize that even though everything is 'hackable' (in both physical in terms of software), it can be useful in slowing down the enemy (which could potentially make them give up), at least enough so you can intervene. At least, this is how I see the use of 'security' or 'deterrents' if you wish.

[Art]

Yeah, and if my security should fail, there's always old Bess, and you really don't want to run into her. Smile! Wait for the Flash!!