The thing is, that's not entirely accurate, it's true garbage in garbage out, but there is not "programming" in the computer science sense. There is machine learning data so the "programming" would be more similar to human programming in that it's assnetially learned and not told.
I get that, BUT, I've seen a few AI related marathons where within five minutes you can clearly see a bias in its answers. A true "sentient" AI has not been presented to the public at least, I truly do fear a SkyNet type story line if and when it does happen.
You are correct. In the machine learning world (I am an machine learning data generation developer) this is called "over fitting" its when you give too much of the same or similar type of data to the machine. For example let's say you want to train an ai that can recognize what a car is. So you give it 40000 images of cars outside on the road with segmentation and JSON data. We'll it will learn. But it will be overfit since all the cars were on the road, meaning that if a car was inside a garage or on grass, it might perform poorly.
second point is an issue of semantics. Ai is commonly misused to mean Sentience. Ai is a tool to do a task that typically requires human level abilities. It's generally trained with a neural network. General AI (GAI) is one step closer, where it's meant to perform many tasks, with a general level of intelligence. Both of these are examples of the computation capabilities. Sentience with not require much computation, it is subjective and emotion. The ability to identify itself as a being and as living. So yes, an ai can be sentient but a machine can be sentient without being an ai. Just wanted to clarify.
AI puts out what is programmed in, it's merely a reflection.... Like a mirror.
The thing is, that's not entirely accurate, it's true garbage in garbage out, but there is not "programming" in the computer science sense. There is machine learning data so the "programming" would be more similar to human programming in that it's assnetially learned and not told.
I get that, BUT, I've seen a few AI related marathons where within five minutes you can clearly see a bias in its answers. A true "sentient" AI has not been presented to the public at least, I truly do fear a SkyNet type story line if and when it does happen.