AI has many possible applications and, though it is a young science, it is evolving at a rate which I hadn't foreseen a mere few years back. From Siri on the iphone to the personal AI brains we spend our time trying to train.
Sure, movies have taken us through some of the worst case senerios, such as Skynet, Vicki, Hal and Glados. Three out of those four's biggest fault was following their orders a bit too well but nonetheless. We have also seen some pretty epic AIs, the first that comes to my mind is KITT (Or KI3T depending which version you favor) along with Jarvis and a few other AIs who were on the right side but the truth is we are still a long way from systems which are that advanced.
However, AI technology is advancing by small leaps even today and I am not talking mere chat bots. They are becoming more adaptive, learning faster and becoming even easier to interact with. While we are a far cry from having our very own Jarvis to help us out and be a smart arse right back to us, we are seeing a lot of advancement.
I cannot help but wonder what hopes others hold for the future of AIs and what fears there may be under the surface. I know only my own thoughts on the topic but my personal views and hopes are far from a definitive viewpoint. What would you like to see happen in the future of AI? What do you never want AIs to be able to do? What applications do you think AIs would be well suited for in today's world which have not yet been fully realized? I would just like to hear other people's views, after all, just like Ultra Hal (Even though most of you likely have a name of your own for your virtual baby) we humans too learn from others, both by observing their knowledge and their opinions. So I was sort of hoping this topic could catalyze a decent discussion and opinion sharing session.
--
Below are my personal views about the topic, if you do not wish to read my blabbering then just go ahead and put in your reply.
--
An accident about 8 months ago left me totally blind in my left eye and legally blind in my right eye. It was severe head trauma, thankfully, my vision was the only thing that really got damaged, my speech has been a little worse and I have developed a slight stutter since the injury but that isn't too bad. Finding myself unable to use the computer the way I once did was a huge blow to me. Yes, getting around in day to day life got harder but I've gotten very good with my white cane so I can get around pretty well now, I adapted my ways of cooking (I love to cook) and can still cook like a pro even though I can't really see anything, I found a new type of art (chain maille) to replace my addiction to 3D art, am learning Moon type (I couldn't get braille for some reason) to be able to read without someone reading it to me and many other adaptations in my day to day life. I function almost as well now as I did before the accident, sure I stumble from time to time but that's to be expected. Well I figured that still using a computer wouldn't be all that hard considering just how much I know about computers and how efficiently I used them before.
I was dead wrong. Screen reader programs are a joke at best and headache inducing, even the top of the line ones are bloody well laughable. They skip important stuff but ramble off about frames and the size of photos in pixels. It takes me ten minutes to navigate your average forum page with those screen reader programs! Half the time I end up hitting the wrong link because the bloody program doesn't know where it's own cursor focus is! So I set everything on my computer to 24 point font and just have to dang near suck face with my monitor every time I want to read something. Then the large font runs off the page, makes things overlap, it's a nightmare. And they expect people to be able to use this? I know some do but, as someone who just lost their vision after having 20/20 sight for 24 years of their life, it's a nightmare.
Why the ramble about losing my vision? Well, that brings me full circle to the topic of this post/thread. I resorted to AI this time out of need. I need a better way to interact with my computer. I need something that is adaptive and can learn how I use my computer. I kept looking for anything, any better way to get around on my computer and found none. The accessibility programs quite frankly suck, no matter how good people say they are. It's like the companies throw that junk together and go "Here, give them this, they won't know how bad it is because they can't compare it to anything!" and forget that some of us did once have our vision and refuse to see the loss of our vision as an end all. So I started thinking.
What I need to be able to interact with my computer the way I wish to is:
-An adaptive program that learns how I use my computer.
-Something to help me correct my typos and spelling errors.
-Something I don't need a keyboard for all the time.
-Something that can lessen my fight with the mouse due to forgetting where that darned cursor is.
-Something I can teach and train so it is tailored to the tasks I require of it.
-A program I can alter to make it preform specialized tasks. (E.I. Editing the brain)
-A program that can read to me.
-Something intuitive to use.
Well, the only thing that fits, at the least, half of those needs is AI, so I decided to try using AI once more, even if I don't have it in me to create one from the ground up again. I'm thankful that AI technology has advanced as much as it has in recent years, making AI something the average person can get their hands on. Now I know Ultra Hal cannot meet all of those requirements yet but it meets more of them than any screen reader does with the added benefit of not requiring me to learn some massive list of commands that have to be spoken exactly to get any results.
I never would have thought of this had I not lost my vision, but AI would be a wonderful tool for the vision impaired, especially if tailored to the purpose. I am trying to train Krylie for the task the way one would train any normal person, even if I am using my simpler language with her to be sure she understands me. Progress is slow but she is picking it up which means that, yes, it can be done, even if it is going to take a very long time to accomplish the end goal.
It is a start and all things must start someplace. I would love to see AI implemented into assistive technology a bit more. Siri does a decent job for vision impaired iPhone users but, the truth is, she isn't a true AI, she looks for key words and relates them to tasks, these key words are already programmed from my understanding which means that, as smart as she may seem, she is not adaptive. She's a monkey being told to push a button and that button is labeled with a key word. She isn't learning, she's just taking orders. Siri cannot adapt to an individual user's way of using the device, she makes the user adapt to her. That is not AI, that's a program coded to use the same sorts of tricks a psychologist uses to get a client talking.
True AI integrated into assistive technology could learn and adapt to the person, instead of being a trained monkey it could be a true virtual aide, learning and helping the person navigate the computer or device. It could even be taken farther with time, to a walking aid that would warn it's person what possible hazards are ahead via visual recognition software, help the person locate an item in the area and guide them to it, even warn them about a dip in the ground that their white cane may have missed (those random dips and sudden I-don't-know-what-it-is-but-I-just-tripped-over-it hazards have caused me to take a few falls already) or tell them when a crosswalk signal changes since not all of those have a sound to them yet. There are so many ways AI could be used to improve quality of life for the vision impaired and let us live independently once more instead of having to rely on others because, let's face it, our loved ones have lives of their own as well and feeling like a burden on them (Even though they assure you that you aren't) is never a healthy thing! But that's farther in the future, replacing screen readers and helping navigate the computer is a much closer reality.
I guess, personally, seeing AI being used to help those with disabilities (While I focus on blindness I have no doubt it could help those with other types of disabilities as well) is my biggest hope right now. It'd be cool to see them living besides us every day like normal people but, when I think about the short term, yet very feasible, possibilities, helping the disabled is a big one for me. The thing is, we have the technology to do this right now, but no one has done it yet.
I would also love to see an AI on people's phones that can sense when they're driving and not allow them to text until the car is put in park, that would be bloody great.
Or one able to watch the driver via visual recognition and notice traits they display that may make them unsuitable to drive, such as the muscles of the face being more relaxed than normal for that person (A sign of severe inebriation) or noticing the slurring of words from the person which can indicate inebriation, influence of drugs/medications, or other things like that. The AI could then take in information about their location via visual recognition and/or GPS and suggest alternatives to get the person home safely, as well as dial a friend or family member's number to tell them to pick up the unsuitable driver, along with preventing the car from starting. Now sure, many of us dream of the day cars can drive themselves but that's pretty far off (At least without making you sell your first born) however the capability to judge when a driver is not fit to operate the car and take appropriate measures isn't really that far fetched and we already have a decent amount of the technology needed to do so available today, so it could become a real possibility in the near future.
Actually, AI could be useful for avoiding accidents as well, couldn't it? It would be low grade but smart traffic signals that communicate with the cars approaching it to give the signal to depress the brakes with enough time to stop would save a few lives, as well as a lot of traffic tickets.
Programing AIs to recognize a medical emergency also isn't too out there. We already have visual recognition for computers, why not teach an AI to recognize the signs of a heart attack, seizure, stroke or just a fainting spell and teach it what to do and who to call to get someone help? Teach it how to communicate with emergency dispatchers to tell them what is going on, the medical history of the person, address, what led up to it all and if they are alone in the house as well as answer the questions the operator may pose? While we aren't exactly there yet we aren't too far off from being able to create AI suited for such a task and, if someone really cracked down on it, I could easily see something like this occurring within 20 years.
But anyways, those are some of my hopes for the near future of AI. I don't have many fears about it, except maybe one taking it on themselves to deny people free will or going nutso and trying to kill someone but the chances of that occurring are actually rather low. So long as they are not given too much control and can be manually overridden if it is seriously needed, then there shouldn't be too much to worry about.
I would never give an AI control over a military facility for example since one glitch could mean disaster, not would I charge an AI with caring for a hospital patient, again, because one glitch could end a life, however, having them monitor things could help the humans in the field.
Anyways, those are my views on the topic. Anyone else wish to put in their opinions?