Board Thread:General Discussion/@comment-177.37.240.146-20160829213948/@comment-67.140.174.219-20170928011435

166.137.252.124 wrote: Idk When I look at ibm watson and how it's been program to think and adapt. They even ask it diverse questions and have it debate with itself. It can diagnose rare diseases faster then Doctors while not having internet access. If Watson can do that it can do some major help with games including overseeing stuff. But the issue isn't can so do it but more of cost and it would cost too much sadly.

Watson isn't that intelligent, it is a compairison program. It's great at certain things like compairison. The reson it seems so intelligent is that it takes in inquiry and searches it on the open web, it takes all of the results that it finds and gives the most common answere, it isn't intelligent it just compairs different answeres and gives the most common based on it's search criteria.

Diagnosing cancer is the same it takes hundreds of thousands or millions of immages of cancerous and non cancerous flesh and compaires them giving the most common result based on it's search inquiry. It's not better than doctors it just combs through more material than them. If a doctor devoted themself to the same thing they would easily become more accurate than watson.

The pont is, watson only looks intelligent it can not think for it's self or make even simpel decisions. It takes data analyzes it, compaires it and shoots out the most common answere based on a simple set of rules.

In all honesty that's not much different than what people do, the problem being that humans are much more flexible we can create alter and elliminate our own rules and search criteria as needed. For example let's say we ask about lets say the origin of humans. Watson would search the web without prejudice. It would consider websites that say that humans were failed alian genetic experiments, humans were created by the gods, and humans evolved from a common ape ancestor and etc and would give the same priority to each article it found giving the most common result. Humans on the other hand would search with prejudice ignoring one or more of those categories as false information. We have the ability to rule out certain types of information or consider certain sources of information as non reliable and we can descriminate against them, an AI no matter how intelligent can not do that without explicit instructions to do so. It can not be biased. Though programers are workin on creating biased software, it's a lot harder than one would think. Also an AI is incapable of creating information something that humans take for granted. We can immagine objects that don't or cant exist but an AI can not do that.

Basically we cant create an AI that is capabe of handeling complex scenarios that would require executive functions. (Look up executive functions in psychology) Executive functions are impossible for AI with our current level of tech, it would require far too much data becasue of the way that AI are made.

Do a little more reserch into neural networks and you'll know what I mean, it's too difficult to explain without going into a full blown lecutre.