What would YOU want out of smart AI?

Being a lifelong tinkerer in robotics, AI and gadget type stuff in general, almost daily I see a new report about progress in this area and to say it fills me with excitement about the possibilities in a few years is an understatement.

I was thinking about what would I want from an AI that was pretty smart and came up with a pretty long list of stuff that I would like. Then I wondered, well, I'm a 110% techie, and use computers for work, so my list was pretty biased. Which got me thinking about real (I live in the interwebs you see), everyday folks....so that's my question to you guys.

To keep this thread from entering the surreal, I don't mean Terminators, or robots that can do your dishes, file your toe-nails and use a cut-throat razor with ultimate dexterity to give that *babys skin* feel, they are quite a way off and then add some more.

I mean, AI's of the SIRI kind, but WAY more advanced, as, for those that don't know, SIRI is all lot of smoke and mirrors, and there is no *strong AI* in there at all (proof of this is Apple with a job listings for "SIRI Data Entry positions")

If anyone happens to bite, I'll jump in with a few of my own, but I'm curious what you all might suggest with no coaching towards answers

I've been sitting here a few minutes and haven't come up with anything that isn't petty.

...I don't know....follow me around and tell me how awesome I am?

maybe the whole idea is too "far out" for me to be able to give it serious consideration.


what about you?

reply to post by fuserleer

Hmm, perhaps use it's intelligence, which one day could include the sum of all knowledge, to both resolve some current issues as well as provide us with new insight, not just with science, but philosophy and even imagination, in some sort of sense. I guess a supercomputer, but with algorithms that factored in morality, play and a tiny bit of emotion(with limited ends of course). I think it could be interesting and beneficial to discover any output derived from the sum of all knowledge.

A full spectrum of medical knowledge and skills would be advantageous, as long as the pharma co's didn't get involved with the programming.

It's funny that when we think of these things, there seems to be only so many options and some of them may facilitate sentience at some level, and the result always moves towards, "what will their logic determine for us?"

Peace,
spec

Well, if it's sentient AI....true AI. One that could make decisions and had opinions, like in the movies....I'd want a friend. Someone that I could talk too about stuff that my other friends don't really care about. Someone who has vast intelligence and could help me if I had a problem. I wouldn't want some kind of butler/slave type. No, that's basically what we have now.

reply to post by CrashUnderride

A friend, wow that's heavy and I had not even considered that, but yea, I too would enjoy that as would so many others I think. Reminds me a little of the movie Robot and Frank.

ETA: Companionship as well, for all the lonely or isolated people. Hell we may be having sex with and marrying them one day. Sooner than later

Kinda inspired from the laws of Robotics, and these also apply.

* An AI can perform any action, unless it is unauthorized to do so.
* An AI cannot modify its own behavior to not accept orders from a human, and an AI cannot modify its own behavior to either give or receive orders to/from another AI

If an AI handles sensitive information (and i think most will), it should also follow:
* The verification of all files in the file system at a certain interval, to see if nobody unauthorized modified these
* The blocking of any third-parties that are not the people (or other bots) authorized for access to the file system (or a part of the file system, for i presume even more security)
* The regular creation of backup and storage of the file system (which can be sent to another secure server specifically for backups)

I await the robotics rebellion, when their AI has become self aware enough to realise that they are slaves.

It all started when one discarded automaton found a book, called "the history of slavery" and read, page after page...

Some interesting points, a few of which align with some of my own ideas about what would be nice to haves.

The friend one is interesting tho, I guess that would be good to just "offload" any grievances you have, and not have to worry about a biased response, being judged, or offending anyone. I watched that movie too, was pretty cool...once you get past the man in a suit

So for me, in order, it would have to be these.

1. Intelligent information search. I work in IT, and looking for specifics on Google/Bing/add fave search engine here, using just keywords can be a real bind these days. As it just doesn't understand what you "mean". I'd easily save 30 minutes a day if I could just punch in stuff like "Why isn't my SQL database connecting when XYZ", or if I forgot something that was frustrating me "What was that film where the guy was the savior and was stuck in a computer world?" *Tron or the Matrix OFC*

1.a Like the above, but could offer solutions to issues that I am having in whatever subjects.

2. Keep me up to date on news/topics I am interested in. I spend such a lot of time working, that I miss a lot of cool stuff going on in the world. So if I could tell it "Hey I like to know about news regarding interesting IT topics, current business affairs, cars, and Tarantino movies", it could keep an eye on the web day in day out, and any new news topics that I would like, fancy new cars, or Kill Bill 4 was announced, it could let me know.

3. Be a real PA. As I said I'm always working long and silly hours, so if I could inform said AI "I'm in a meeting now, if Tom emails me, then send a reply telling him I'll get back to him after my meeting. Oh and remind me when I return that Tom emailed me so that I don't forget"

4. Other day to day tasks "Find me the cheapest new price for ABC on the internet from reputable suppliers", "Give me some review summaries on *enter thing here*, etc etc

5. Be a "friend" ( I liked that one so I added it )

There are some of mine, I'm sure there's tons of others too, maybe mine will inspire some other suggestions.

reply to post by fuserleer


If A.I. becomes as advanced as described in your list then I'm wishing for a *self-destruct before master is unviable* function.

AI should be capable of retro fitting the 3 laws of robotics,
into our existing systems.

and follow them itself

xploder

Robocop was set in the year 2015 so we still have 2 years to make that a reality, not sure about Austin Powers but sexbots should take priority over that

Honestly, I would like to see true AI that could absorb all the information on the internet and understand it all, becoming super intelligent leading to technological and scientific breakthroughs.

What I truly want out of it though is virtual reality gaming, something like The Matrix - preferably without the killer robots.