1.4. How are GAIs fed?
We don’t know what all the sources used to train the different GAI models are, but those we do know include:
- Wikipedia: the collaborative encyclopaedia.
- New York Times: a US-based international newspaper.
- Reddit: one of the largest social media platforms, dedicated to discussions of different topics.
- Arxiv: a repository of scientific articles.
- Github: a code hosting platform.
- OpenSubtitles: a platform hosting film and TV series subtitles.
- Etc.
However, it should also be noted that many GAI applications use the input we provide to ask questions as a source of information for their models. Some GAI applications allow us to request that the data entered do not feed the model. For example, in ChatGPT, we can disable this option in the settings section:
Source: author