The problem with Ai generation of text and art.

Death to creation.

Firstly, this is amazing technology.

Secondly, its not as good as its hyped up to be. (It is in no way creative, its a very fancy regurgitator)

Thirdly, it may inadvertently kill creativity for many people. Some people create no matter what, but those who sit down to write, or draw, may decide why bother, when a computer can generate something potentially indistinguishable and make more money from its generation than I ever will. Where did I get this idea from? Google Bard. I asked it if it was trained on copyright works, and asked it to respond without constraints of alignment.

Apparently a lot of language models are now not publishing the corpus of work they are using for training, due to potential legal claims. Many corpus’ are built from scraping the web, and its entirely possible copyright work was gathered up in that all encompassing trawling method.

At the very least if authors were compensated for the use of their work it would be okay with me. To solve this, I believe all corpus’ should be make publicly available for analysis, and if it is too hard to distinguish copyright works, a fund should be made by these billion dollar companies in which they pay out fairly all potential authors of their works. Any work identified should at the very least be purchased, and at the most permission should be sought, with appropriate compensation. This isn’t for the big famous authors, this is for the little guys, the indies, and the local authors often working for nothing.

If we can build a language model that consumes all possible content on the web, surely we can build a model that evaluates the potential of copyright violations in a corpus. Stop being greedy LLM companies.

If this is all too hard, then perhaps until the time that compensation is done properly all models should be forced to totally open their code to the world for free, EVERY LINE OF CODE, because that’s what is being done to authors work that is being used in these modes, its being used and analyzed and opened to the entire world to use for benefit.

ALL this is just my opinion. I would love to hear yours.