2026-04-23 00:01 UTC
@jezebelkat@famichiki.jp Having a policy on LLM risks legitimising it. Unless it is absolute, it provides a foothold. Imagine a policy that said "You may consult with Nazis, and consider their advice". Or "You may hire Nazis to do some of your work". Or "by all means employ Nazis, but make sure your check their work carefully".
The only policy that makes any sense with generative AI is one that says it is not to be used, at all, by any person at any level, in the processes and procedures of the organisation. And that care should be taken to avoid inadvertently using, passing on or being influenced by, material from outside the organisation that has been produced with the aid of generative IT.
Put the reasons in an appendix or in footnotes, not in the policy.
Replies (1)
-
@jezebelkat@famichiki.jp 2026-04-23 00:11
@kauer@aus.social That is EXACTLY the plan for the draft policy, as stated in the original toot. The starting position was going to be "no, none, not at all, never" but as we are a democracy, and one with over 500 members, we cannot guarantee that the ending policy will be no use at all, but by starting at no use at all, we hope that any movement to use will be heavily constrained. But we are a bottom up organisation. We will not impose policy.