Generative AI: Difference between revisions

Line 35:

https://www.deviantart.com/team/gallery -->

=== LAION-5b training database ===

===LAION-5b training database===

Many users have had their content scraped by LAION to power their training database, and the only way they can opt out is via a third party<ref>https://haveibeentrained.com/</ref>.

==== <big>STACKOVERFLOW training data</big> ====

====<big>STACKOVERFLOW training data</big>====

In mid 2023, StackOverflow released [https://stackoverflow.blog/2023/07/27/announcing-overflowai/ Overflow AI] which uses all users questions and answers for their neural network that they charge enterprises for. Despite their strong stance on [https://stackoverflow.com/help/gen-ai-policy Generated AI Post Policy] they still require users to not generate AI responses despite scraping your human content for profit. As of now there still is no official way to opt out officially other than deleting all your posts/topics manually before permanently deleting your account. But even then their [https://stackoverflow.com/help/delete-content Get Out Clause] still allows them to use it even after you thought you deleted it. They have completely burned their bridge of privacy and user trust for the platform and brand and the users content that the site was built by.

=== Microsofts Github Copilot training on free user repos ===

According to this [https://copilot.github.trust.page/faq?s=v2qe7voltpwtv2usl4ikhs#ip-and-open-source Copilot FAQ] topic it will not train on people that are on PRO or ENTERPRISE plans but says nothing about the people using Github under a free account. With that being said we can assume that any and all accounts that copilot is in currently is being used to further train the model. As creators credit is the least you could do; especially if your profiting off work that is not in fact yours. Your platform is supposed to give people an alternative to price tag software and the passionate hobbyists that modify, scrutinize, and improve a project for the community. Instead you have eroded [https://github.com/orgs/community/discussions/152229 Users Trust] in your platform by choosing profits over people.

==References==

@@ Line 35: / Line 35: @@
 https://www.deviantart.com/team/gallery --><!-- Due to my close familiarity with the situation, yes, I developed this section a lot more than initially planned. -->
+=== LAION-5b training database ===
-===LAION-5b training database===
 Many users have had their content scraped by LAION to power their training database, and the only way they can opt out is via a third party<ref>https://haveibeentrained.com/</ref>.
-==== <big>STACKOVERFLOW training data</big> ====
+====<big>STACKOVERFLOW training data</big>====
 In mid 2023, StackOverflow released [https://stackoverflow.blog/2023/07/27/announcing-overflowai/ Overflow AI] which uses all users questions and answers for their neural network that they charge enterprises for. Despite their strong stance on [https://stackoverflow.com/help/gen-ai-policy Generated AI Post Policy] they still require users to not generate AI responses despite scraping your human content for profit. As of now there still is no official way to opt out officially other than deleting all your posts/topics manually before permanently deleting your account. But even then their [https://stackoverflow.com/help/delete-content Get Out Clause] still allows them to use it even after you thought you deleted it. They have completely burned their bridge of privacy and user trust for the platform and brand and the users content that the site was built by.
+=== Microsofts Github Copilot training on free user repos ===
+According to this [https://copilot.github.trust.page/faq?s=v2qe7voltpwtv2usl4ikhs#ip-and-open-source Copilot FAQ] topic it will not train on people that are on PRO or ENTERPRISE plans but says nothing about the people using Github under a free account. With that being said we can assume that any and all accounts that copilot is in currently is being used to further train the model. As creators credit is the least you could do; especially if your profiting off work that is not in fact yours. Your platform is supposed to give people an alternative to price tag software and the passionate hobbyists that modify, scrutinize, and improve a project for the community. Instead you have eroded [https://github.com/orgs/community/discussions/152229 Users Trust] in your platform by choosing profits over people.
 ==References==
 <references />