ChatGPT is trained using data up to the year 2021. Therefore the answers given are sometimes less accurate and may be out of date. To solve this issue, OpenAI produced GPTBot that searches the internet to get the latest information. But ethical issues arise when data from non-OpenAI sites are used without permission to train ChatGPT.
In order to prevent this issue from recurring and also to prevent themselves from facing more intellectual property abuse lawsuits, OpenAI is now empowering site owners to block GPTBot. By including a Robots.txt text file or blocking IP addresses, GPTBot can no longer read content on pages without permission.
The use of open data to train artificial intelligence has become a hot issue in the last year. It's one of the reasons why Elon Musk imposed a temporary ban on Twitter because he claimed user data was being used to train AI. In a lawsuit filed in the United States, several authors claim that OpenAI used thousands of books to train ChatGPT without compensation.
Tags
APPS & GAMES