The
Wikimedia Foundation
has announced,
on the mailing list
and
on Meta
, that there will shortly be another vote on the Enforcement Guidelines for the Universal Code of Conduct:
In mid-January 2023, the
Enforcement Guidelines
for the
Universal Code of Conduct
will undergo a second community-wide ratification vote. This follows
the March 2022 vote
, which resulted in a majority of voters supporting the Enforcement Guidelines. During the vote, participants helped highlight important community concerns. The Board’s
Community Affairs Committee
requested that these areas of concern be reviewed.
The volunteer-led
Revisions Committee
worked hard reviewing community input and making changes. They updated areas of concern, such as training and affirmation requirements, privacy and transparency in the process, and readability and translatability of the document itself.
The revised Enforcement Guidelines can be viewed
here
, and a comparison of changes can be found
here
.
Voting will open on
17 January
. For information on how to vote, eligibility, and the like, see the detailed
Voter information
page on Meta.
?
AK
WMF General Counsel
Amanda Keton
has
announced
that she will move on to her "next adventure". She joined the Wikimedia Foundation in 2019, having previously served as head of the
Tides Foundation
and CEO of
Tides Advocacy
.
Amanda will be succeeded in her role as General Counsel by current Deputy General Counsel Stephen LaPorte (also known as
User:Slaporte
), who will take over on 1 February 2023.
The Signpost
wishes Amanda and Stephen all the best. ?
AK
Large language models on Wikipedia: friends or foes?
- Note: JPxG, who wrote some of this section, also wrote
WP:LLM
.
For the last few weeks, a
discussion
at the Village Pump has been ongoing about the potential use of text generated by large language models (like
GPT-2
,
GPT-3
,
GPT-J
and
ChatGPT
) in Wikipedia; near the end of December, a
thread on the wikimedia-l mailing list
discussed the issue as well, going over the potential benefits, drawbacks and use cases for these models.
Recent
coverage
in
Slate
describes a series of discussions which took place at
Talk:Artwork title
concerning the article (
Artwork title
), whose initial draft consisted of prompted output from
ChatGPT
. Various demonstrations have been done of the use of these models to assist in writing, editing and formatting (
User:JPxG/LLM demonstration
and later
User:Fuzheado/ChatGPT
). Currently, a
proposed set of guidelines
for the use of these models is under discussion (and, frankly, could use some more eyes on it).
The August 2022 issue of
The Signpost
explored some of these issues as they concern this publication; see the
From the editors
page titled
"Rise of the machines, or something"
for an introduction. ?
J
,
B
Brief notes