So what can Artificial Intelligence (AI) do and what can't it do?
We have AI already around us kind of all the time, managing all sorts of different things. There are traffic control systems that manage airplanes like auto landing control or ATC (Air traffic control), The auto landing control systems help tell if a plane is too low or too high and help it land. Autopilots are another example of airline AI, They can determine any number of different things all at once there are certain airplanes that fly by wire and the plane is always in control. The pilot can give input to the autopilot but there are manufacturers who believe that the autopilot is safer than the pilot so, if the pilot decides to turn too far the autopilot won't let them turn more than certain number of degrees if they are have passengers on board.
There's also autonomous driving cars. They're reading the streets they're reading the surroundings they're reading in real-time stuff that's happening around them all over the place. Driving is a very stressful situation for the human brain, we are terrible at it but for a computer it's bringing in all of these variables and figuring it out. And on top of that the reason you could call it AI is because there is that variable of human pedestrians, I could jump in front of an autonomous car and that car has to make a decision does it turn, does it stop, can it accelerate around me, how is it going to deal with this situation if that's not thinking I mean yes it's programmed thought but it's still there.
Siri is artificial intelligence, Cortana, google those services are all a rudimentary artificial intelligence. They're listening to what you're saying and there are attempting to determine the answer you're looking for and they're doing it based on context clues. You can tell Siri that you have a mother whose name is this and it will remember that and be able to respond and you just say text mom even though you identified her as mother these services will be able to tell the difference. They understand that mom, mother and mommy and all of these other little language idioms are all the same entity which it's been programmed to do again which is why it's a rudimentary AI. Voice recognition in general is sort of like an AI system.
There's also modern AI which is UAVs and drones, we mentioned self-driving cars that's kind of some of the newest AI we have around. UAVs and drones are really interesting because they're designed to be artificially intelligent in that . They can monitor wind speeds they can adjust turbulence, they're flying at certain altitudes and in certain ways is to minimize and military capacities minimize enemy detection so they're either flying very high flying around flying very low things that human brains have trouble with is flying too low to the ground you know if things change we might overreact and we could crash the system. The artificially intelligent drone doesn't have that problem, they're programmed to do any number of things based on input, so they can look around they can get to an area and flying in a circle and that's all they have to do, they can get to an area and take pictures of certain predetermined areas, they can also look for specific things like faces or specific models of vehicles and then take pictures of those. Again programmed intelligence but definitely artificial intelligence they are not thinking necessarily but they are making decisions often they are piloted as well, the pilots usually do the more militaristic tasks like actually dropping bombs or something.
We're constantly surrounded by minor artificial intelligence is really what I'm trying to get at. They call this in the artificial intelligence community I guess week AI, their narrow tasks they're very focused they can only do this one thing that can only make these few decisions like a self-driving car cannot drive anything but a car can't fly a plane now if it could that be pretty amazing but it just drives cars. Other examples of weak AI are things like a roomba which is a cutest little robot it's clean your little room for you and then go back to the little dock they know where their home is, they know where your furniture is, they know where the edges of what you want clean or what you don't, they know where they've cleaned in the past and where they should clean in the future. This is not an ad for rumba, I'm just saying they're funny they're kind of smart.
The GPS system is artificially intelligent in a lot of ways or pretty much anything that uses it because they can understand where it is and where it needs to go. Recommendation engines like Amazon and Netflix they know your past behavior and they predict your future behavior based on a variety of algorithms, that determine what someone bought when they looked at this specific thing how many other people bought this other thing and they can suggest that to you. But they can suggest it in different ways maybe you go to look at a DVD and it knows maybe I'm bought a DVD player so i can recommend a DVD player or a new player or that knows that you've looked at audio equipment in the past and now you're looking at a DVD player hey guess what you need this new blu-ray because you've looked at this other blu-ray thing before, so I'm going to recommend these blu rays that you should get with your new blu-ray player it's making decisions that is weak AI.
The thing is though that is cool as all that sounds, I mean it sounds like Amazon knows what's up but weak AI sucks guys, It's the worst it's really really really bad. Mainly because computers are dumb they only know what we tell them they don't create new thoughts on their own. So image processing is an excellent example if i show a computer a picture of a sunset it doesn't know what to do with that picture, it can analyze it for the colors and what is perceived to depth and because it's been programmed to know what depth might be in that photo and what the colors might be but it's taking the camera the software and then a huge amount of programmed knowledge bases. So lots of other photos that it looked at and been told this is the depth part, this is the pretty part of the sunset, this is what ground looks like, this is what water looks like. It uses all of these things and then when we look at it we see a sunset it sees a variety of data and then doesn't know necessarily what to do with it they don't just look for the interesting parts of the picture they aren't looking for the sunset and their look oh my gosh there's a little person standing on the edge of this cliff in front of the sunset. That's the cool part the fact that it's this huge vista and there's this tiny little person it would never even know it would look at every pixel the same. Human brains look at that image and they find the things that are interesting the sunset the water the cliff and the tiny little person, but the pixel next to that person and the person's head are the same to a computer it doesn't know the difference.
Conversational language processing is another thing that weak Ai sucks that we talked about chatbots before. The reason that chat box have gotten so much better the reason that google voice search is really good is because back in 2007 they created something called 1 800 or maybe 188 goog-411 don't dial it not a thing anymore close in 2010. I actually used to use this when I had a flip phone, I'm that big of an neard i would call Google and it would answer and we could play a little tone and you would tell it what you wanted to search for on google and it would read back to you three different results and then you could hang up and you could answer your Bar trivia me like I was my mom and then i could get the answer but that doesn't exist anymore. But from that they use those voice snippets does not like radio we're just disappears they're saving all of that voice input and they use that to create what's called a phoneme database. Phonemes are little snippets of conversation little snippets of voice that they could use and analyze and determine what people were saying. It got really good at understanding voice inputs that's how google built their first voice input database.
A company called Nuance Communications does that too using dragon speak which is another company nuanced I don't know if they do dragon i can't remember but Nuance Communications was a company that powered an app called Siri which was just an app on the App Store you could download I had that one too. Then apple bought the app removed in the app store and there is some debate as to whether Nuance Communications still runs it that, I don't think they're really allowed to talk about it. But they're phoneme databases also really really good so they wanted Siri to keep being powered by nuanced because it could understand what humans were saying. But not unlike image processing it's not really understanding what your saying it's not listening to you say words it's listening to patterns, it's listening to every little piece of information and looking for the patterns that it's been taught to look for.
The reason that when you're online you have to type in numbers and letters obscured by lines is because a computer can't tell the difference between that and any other part of that web page. As far as the computer is concerned all those are just pixels and those lines make it in unintelligible for a computer. But for a human they can look at that and they can say oh this is the house number its this number I'll type it in or this is the word you know neuron covered in some weird colors I'll just type in neuron and I'll keep moving, but a computer can't tell the difference computers are so dumb that what happens is they call it a leap second, where they'll add a second part of a second at some point of the year and when that happens the computer sees it as the same second repeating so it's you know 1 59 59 59 now it's 1 59 59 59 again and it doesn't know what to do so the computer starts to freak out and it's like I already did that second and then it overloads and the computer crashes. This was the whole basis for y2k that's how stupid computers are they can't tell the difference between one second and the next second whereas humans may not even notice. We haven't even gotten into the complicated things all we've talked about is looking at pretty pictures listening to people talk and being able to talk back and understanding that time is meaningless and we haven't gotten to love or emotions we have gotten to faith not to mention the computers don't even understand any of that. Emotions are complex interactions of a number of very discreet hormones in the human body and it uses conscious and unconscious systems you can't even make a self-driving car go faster and 25 miles an hour and gets confused.
If you want to know what the hell is this AI thing then check my post below. https://steemit.com/artificial-intelligence/@masonmiler/what-is-artificial-intelligence-ai-will-artificial-intelligence-surpass-human-intelligence
Image Credits - 1 2 3 4 5 6 7 8 9 10
If you enjoyed, Resteem and Follow me @masonmiler for more content just like this, and as always, take care of yourselves, and take care of each other :)