|
There is a lot published about AI during the last months, also for the general public, but there is an aspect about which you don't hear a lot.
Sabine Hossenfelder (see https://youtu.be/jKx0XwJAj3k) talks about the different USES of AI, and lots of video's and articles do so, but nobody talks about the different ways which are used to make programs act "artificially intelligent".
I suspect there are lots of different types of AI, and the implications for life, society and work might depend very much on the types of AI. Some types of AI give the impression to independently take decisions, as if they think and act for themselves.
Others, however, only seem to improve work already done by humans, like the picture of the black hole improved by AI. (see dr. Becky's video https://youtu.be/RF3P1W1vaYs ).
Others might only be uses to give you lists of recommended products, like video's (YT), but and those might be applications of very simple statistics and links by themes and tags.
But to be able to make those distinctions, one has to know how the so called "black boxes" of AI systems actually works, and that's usually a trade secret. The businesses developping AI are not going to tell it, especially if they have invested billions of dollars in it. And it is the mechanism of AI which might be dangerous or not.
|
