This post does not constitute a final decision, but is a proposal on wording, and call for input on, a suggested stance from the WordPress.org The community site where WordPress code is created and shared by the users. This is where you can download the source code for WordPress core, plugins and themes as well as the central location for community conversations and organization. https://wordpress.org/ support community on how AI-generated resources or replies can and should be used across the WordPress.org support platform.
During the support team meeting on May 11th, the topic on using generative AI in replies was chosen as the meetings discussion topic.
Based on the pros and cons brought forth, and after observing how AI tools are currently being used – not only within the WordPress.org support forums, but also in support-focused scenarios in general – it’s clear that these kind of tools are both quite popular, but also at times misleading.
Some of the concerns raised were that generative replies may not always be accurate, leaving users with an incorrect path forward and more confused than helped, or used to rapidly increase a user’s post count for whatever reason fits their scenario. Another concern is that the replies may be too generalized or outdated. This counts true as tools such as, for example; ChatGPT, are based on older content (in the example of ChatGPT this is 2+ years old) at this time, and its answers may therefore no longer be entirely correct or relevant.
Taking all those elements into account, the following is a short, but applicable suggestion on what our stance may be:
The use of AI-tools to improve replies in support workflows is allowed, but it is on the user to verify that their replies are accurate, and correct. In addition, it is also a requirement that the use of such tools be declared within the response, to provide transparency, and set expectations. Abuse of such tools to provide multiple inaccurate replies, intentional or not, for the sake of activity will be considered spam, and the users will be warned, and subsequent misuse will be treated as any other account producing spam-focused content.
Thank you to @zoonini and @mrfoxtalbot for reviewing and helping with this post.
