Friday, July 10, 2020

A.I. VS Regular Program

Regular programs are on their way out and in its wake will come A.I.. After all, why wouldn't you want every facet of technology to be able to learn and determine the best courses of action for each scenario? Why wouldn't you want a piece of software that becomes better, without changing the code, over time?

The first thing that is important for us to understand is the difference between a "regular" program and an A.I.. A program is code with a verify specific purpose. The code will never do things it wasn't design to be doing. 
The A.I. also is designed with a purpose in mind but it gets to make its own decisions based on its experience. The more experience it as, the better it will be at making the decisions related to his purpose.

The biggest strength, and also the biggest weakness, of A.I. is in its data set. The bigger it is, the more scenarios it has, the better it becomes at the task at hand.

How much data does an A.I. require in order to become proficient at something depends on what that "something" is. Alphago, the A.I. developed by Deepmind to play the Go strategy game, had hundred of thousands of professional games in its data set before it could start beating the top professionals Go players Xray4all, an A.I. designed to figure out diseases from X Ray pictures, only needed a few thousand X Rays of a disease before it could start identifying it in new pictures.

Some A.I. companies, like Deepmind, have found ways to create A.I. that are not dependent on existing dataset. AlphaGo Zero is the successor of AlphaGo and learned the Go game via trial and errors instead of through analysis of thousands of games.

The problem with the "rise of the A.I." is that changes are occurring faster than we can process it. Those technological companies are building A.I. that doesn't consider a user's privacy and some of them are just now paying for it. As I mentioned before, the more data an A.I. has the better it is and there doesn't seem to be much limit of the type of data those big tech companies are allowed to store on us.

While many countries are in on the A.I. race I think it's safe to say that countries with a Dictatorships form of Government, such as China, have an edge over countries where the populace have rights. The reason being is that privacy policies, in Countries like China, are non-existence which gives tech companies the green-light to do what ever they want with the data they gather and they can accumulate a lot of it.

The future is in A.I. and I just hope that we won't become one of those dystopian societies so often depicted in science fiction.

No comments:

Post a Comment