I have not tried this, but there is a suggestion here
that ChatGPT can reproduce material from a copyrighted book.
Books were definitely included in their information intake, but I wonder how far they went. I bet they did not access older books that are only available in libraries or collections. If that is so, their information is biased toward modern material.
There was a Google project many years ago to photocopy every book and make them freely available online. They were stopped by copyright laws. It never happened. The world still has very poor access to its older literary material.
This is a major issue. If the repository includes only Western contents, it is going to give a certain bias. We all know from school that every single nation or culture tends to brush out the worse things it did throughout history. This is a topic i have been working on and would like to exchange on privately.
There is now even an AI business model (sourcely) offering to dig directly in pdf databases. I bet this was without paying anything to the journald. Same way jstor is now requesting EU universities to pay an abo in order to access a digitized version of contents that were actually created by some of these EU universities and with public money.
That sort of thing is ruining public access to information.
When first I worked in research, anything CSIRO created was considered public property and was freely available to anyone who asked.
Then the âintellectual propertyâ brigade took over and in the end I was not allowed to tell anyone about what we were doing, because it might have a âvalueâ⌠not even the people who had paid us to do the research.
Another example. There was a microbiologist at our site doing work on rumen fungi which was paid for by a private company. The results did not turn out as the company expected. They dirested our scientist to not publish the work and to destroy his data. !
Academic freedom is under threat. Universities have fought hard to retain it⌠but wherever there is money involved it can be overruled.
Vested interest is another problem . I once produced a research result that went against a large publicly funded breeding program. The fact that the publicly funded program had a fault was ignored, and I was directed to cease speaking about my work.
My Chief backed me up and refused to interfere with my academic freedom. ⌠but⌠when my Chief retired, they came back and influenced the new Chief to shut me down.
That sort of thing eventually persuaded me to retire from research. It was pointless. I wanted to do something I could be honest about.
Is there any for of political correctness built-in to the AI ?
The subjects we try to avoid such as race religion sex money etc
Not to mention control of dangerous areas
Not given it any thoughts up to now but it not really something I use although I will check the net or YouTube before starting a new project that I know little about. Then use my judgement to decide correct or not
Some AI are built with a certain idea of political correctness. For ex. I tried âwhat are the sources of the old testamentâ and saw that Grok and Chatgpt refrain from mentioning the Gilgamesh epic (unlike Mistral and Deepseek). Musk said that Grok is meant to be a ânon-wokeâ alternative to other bots. Luckily, the multiplicity of free bots might prevent ChatGPT and its Win11 support to conquer the whole worldâs brains or what remains of it.
deepseek is great to give you summaries and outlines on topics which are already well treated in encyclopaedias etc. Tried âoutlines of Platoâs philosophy for beginnersâ âoutlines of Aristotleâs philosophy for beginnersâ. It even worked to search an online database of Greek texts (without even making the request in Greek). More and more people are using the bots as browsers.