ChatGPT is just an algorithm with a huge database and a trained neural network, which by itself is basically a statistical treatment.
There is no intelligence or self-awareness.
Also, all that ChatGPT and Dall-E produce is based on content created by humans.
Truth is a human concept that often depends on the point of view. What is true for one person might ne false for another one. What might be true today, could be debunked as false tomorrow, including science.
ChatGPT is not aware on what truth is. It calculates the weights of the input words and calculates the most probable first output word. Then it uses the input and output words to calculate which is the best next word. This is repeated until the generated text reaches the desired length. This is of course just a simplification. But all there is is a calculation that determines the next word in a sequence of words. No intelligence or self-awareness.
For me, the disruptiveness of this technology is that it accelerates and automates content creation. At the same time it allows to learn new things in less time, is used correctly.
You can not really define what is self aware. If we would suppose something is self aware because it makes unexpected things, than a wrong washing machine would fit the definition as well.
And even if something would tell you that it become self aware, how do you know it has?
And if it has but it didn't tell you?
But the thing is, as we are discussing it now, that even a not self aware thing can do a lot of harm. Just see the bubbling algorythm of FB. They are pretty easy to understand, but the effects on society was unpredictable. (Well actually not, but no one gave a shit I suppose.)
Other problem is, that the methods how these AIs are trained is completely out of any democratic oversight. The natural one become a nazi, so I suppose there is already some taming and modification about what it HAS to think or ansver. And there you have a censorship without any oversight, just like with FB, which is also not solved yet, the algorythms are still not openly aviable.
(How many people and politician would understand them is another question.)
In the last few years I repeatedly asked myself, how issues which affect less than one person out of thousands could become mainstream in the media and society.
While other affecting way more people are almost completely hidden. Did someone fiddle with some algorythm?
So oversiht would be crucial.
PS.:There will be soon an EU US high level political meeting about AI. Will it get at least one tenth of the media coverage like some BS recently?
Home
Help
Search
Login
Register
