Does ChatGPT have licenses to give out information that it does? Even then would it be ethical?
Humans have the right to consume information available on WWW even if content is copyrighted, but does ChatGPT have it? I asked ChatGPT and I am posting a screenshot of the answer received here, but I am far from convinced:
As an AI language model, ChatGPT was trained on a vast amount of text data from the internet, but it is hard to believe that permissions were obtained from everyone to reproduce their content using ChatGPT !!
Let me explain the ethical issues of this using a few real life scenarios:
Scenario 1:
Assume that I am a cook & I post recipes on my own website or my employer’s. When you browse through it, I get credit & may also get business because you may buy from me or my employer. I get neither credit nor business when ChatGPT copies my recipe and gives it to you.
Scenario 2:
Assume that I am a content creator & I post articles or videos on Youtube. When you watch it, I get credit & get advt. revenue and so does Youtube. I may also get business and my popularity increases. None of these are realised when ChatGPT gives out info from what I posted !!
Scenario 3:
Assume that I am running a business from years of experience and I post articles about it from time to time so users get free information, but read it from my website so they may place an order on me. They may also share it with their friends.
Scenario 4:
We humans pay money to get trained, say from Udemy or Universities, even if indirectly in the form of advertisements or increasing their popularity, while browsing through training material. Did ChatGPT make payment or give credit for itself to get trained on these material?
Scenario 5:
Assume I am a developer asking a question on StackOverflow to solve a difficult problem. The question and answers are voted. Everyone gets recognition and may even use it for self-promotion. None of these happen when ChatGPT takes out info from SO and reposts it nicely.
As I understand the law, information posted online is automatically copyrighted and unless a open license is mentioned, the content is supposed to be closed source and requires permission from the author to reproduce it or reassemble it in any form.
While ChatGPT does not reproduce the information as it is, it reassembles it and presents it to the user in a nice way. References: https://en.wikipedia.org/wiki/Copyright https://en.wikipedia.org/wiki/Copyright_notice…
I don’t see anyplace where ChatGPT (3.5) gives credit or shows where it picked up the information.
If I use the Chat option in Bing, it lists the sources at the end, but since the language is so convincing and if the info is believable I don’t find myself or expect anyone for that matter following the links, which defeats the purpose of the existence of those pages.
Most professional websites have fine print that mentioned that the information is not supposed to be reproduced in any form without prior permission from the author(s).
ChatGPT may be within legal limits, but it seems to me that ChatGPT freely reassembles information posted by authors who have painstakingly created content over several years and not even giving them due credit to say the least !!
If you concur with me please pass this on so we can get this clarified! If not, your opinion is most welcome!!
