Oracle has prohibited the acceptance of changes generated through AI tools into the open OpenJDK project. The ban encompasses source code, text materials, and images transmitted in the project's repositories or posted on the website, in messages, or on wiki pages. Areas where the use of AI is allowed for Java development include debugging and code reviews.
The reason cited is the potential risks of infringing on others' intellectual property when using code generated by large language models. The legal status of such code has not yet been definitively established, and there may be potential implications for property rights and licenses under which projects used to train the model were distributed. Security-related risks are also noted.
Additionally, it's worth mentioning the establishment of rules regarding AI in the NetworkManager project. Changes generated through AI are allowed in the NetworkManager repository, provided that the authors take responsibility for the submitted patches, understand the changes, can discuss the essence, have conducted manual reviews, and have performed testing, as well as being prepared to confirm that the submitted code is compatible with the LGPLv2.1 license.
Source: opennet.ru
