User talk:Daniel: Difference between revisions

Browse history interactively ← Previous editContent deleted Content addedVisual WikitextInline

Revision as of 20:39, 25 December 2024 editJesuspaul502 (talk \| contribs)177 edits →Question About Formal Tone and AI Concerns: new sectionTags: Mobile edit Mobile web edit New topic ← Previous edit		Latest revision as of 06:42, 26 December 2024 edit undoDaniel (talk \| contribs)Autopatrolled, Checkusers, Oversighters, Administrators75,520 edits →AFDs: r
(9 intermediate revisions by 4 users not shown)
Line 24:		Line 24:

	Thanks for your input! ] (]) 20:39, 25 December 2024 (UTC)		Thanks for your input! ] (]) 20:39, 25 December 2024 (UTC)
			:Two (out of two I checked) of your posts came up as 90% AI-written according to the AI detector I used (which is falliable, but that's such a coincidence). Further, the subheadings that you used were distinctive of AI-generated messages. I still believe you were using AI to generate those two messages. (By contrast, this message registers 0%.) ] (]) 22:22, 25 December 2024 (UTC)
			::Also, . ] (]) 22:23, 25 December 2024 (UTC)
			:::What's the current guidance on blocking AI trolling, which this clearly is? ] (]) 23:04, 25 December 2024 (UTC)
			::::I don't feel good about feeding back the "tells" either ("subheadings") ] (]) 23:06, 25 December 2024 (UTC)
			:::::There's no automatic prohibition in project-space, so it needs to be judged through the lens of general disruption. The use of subheadings is both a tell but also part of the disruptive nature of LLM-generated projectspace postings, so I don't think it's fair to say "you're being disruptive" but not explain ''how'' it's disruptive. ] (]) 00:18, 26 December 2024 (UTC)
			::::::In response to your fairness objection, I think it's wrong to knowingly assist in calibration of such models. We don't feed trolls, we shouldn't tutor AIs. ] (]) 00:49, 26 December 2024 (UTC)
			\|}

			==AFDs==
			Hello, Daniel,

			I know you needed to take a break from admin responsibilities and I probably feel like that every week! But it's great to see you back, helping out in AFDLand. You've been missed in our regular rotations of AFD closers. Glad to have you back! <span style="font-family:Papyrus; color:#800080;">]</span> <sup style="font-family: Times New Roman; color: #006400;">] ]</sup> 06:33, 26 December 2024 (UTC)
			:Thanks Liz! Yesterday and today have been reasonably quiet for me so thought I'd jump in and help out. Life gets busy again tomorrow for a few weeks but I'll do whatever I can to help! Cheers, ] (]) 06:42, 26 December 2024 (UTC)

Latest revision as of 06:42, 26 December 2024

Archives edit

101 · 102 · 103 · 104 · 105 · 106 · 107 · 108 · 109 · 110

111 · 112 · 113 · 114 · 115

The Signpost: 24 December 2024

News and notes: Responsibilities and liabilities as a "Very Large Online Platform"

Op-ed: Beeblebrox on Wikipediocracy, the Committee, and everything

Opinion: Graham87 on being the first-ever administrator recall subject

In the media: Delhi High Court considers Caravan and Ken for evaluating the ANI vs. WMF case

From the archives: Where to draw the line in reporting?

Recent research: "Misplaced Pages editors are quite prosocial", but those motivated by "social image" may put quantity over quality

Humour: Backlash over Santa Claus' Misplaced Pages article intensifies

Gallery: A feast of holidays and carols

Traffic report: Was a long and dark December

* Read this Signpost in full * Single-page * Unsubscribe * MediaWiki message delivery (talk) 00:00, 25 December 2024 (UTC)

Question About Formal Tone and AI Concerns

Hi Daniel,

I just wanted to ask, does using a formal tone in my responses mean I'm using AI? I’ve been trying to keep my responses clear and respectful, but I didn’t realize that could be seen as AI-generated.

Is it a problem to write in a more formal style here, or does that automatically raise concerns about using AI? I'm just trying to understand what's expected.

Thanks for your input! JESUS (talk) 20:39, 25 December 2024 (UTC)

Two (out of two I checked) of your posts came up as 90% AI-written according to the AI detector I used (which is falliable, but that's such a coincidence). Further, the subheadings that you used were distinctive of AI-generated messages. I still believe you were using AI to generate those two messages. (By contrast, this message registers 0%.) Daniel (talk) 22:22, 25 December 2024 (UTC)

Also, this. Daniel (talk) 22:23, 25 December 2024 (UTC)

What's the current guidance on blocking AI trolling, which this clearly is? BusterD (talk) 23:04, 25 December 2024 (UTC)

I don't feel good about feeding back the "tells" either ("subheadings") BusterD (talk) 23:06, 25 December 2024 (UTC)

There's no automatic prohibition in project-space, so it needs to be judged through the lens of general disruption. The use of subheadings is both a tell but also part of the disruptive nature of LLM-generated projectspace postings, so I don't think it's fair to say "you're being disruptive" but not explain how it's disruptive. Daniel (talk) 00:18, 26 December 2024 (UTC)

In response to your fairness objection, I think it's wrong to knowingly assist in calibration of such models. We don't feed trolls, we shouldn't tutor AIs. BusterD (talk) 00:49, 26 December 2024 (UTC)

AFDs

Hello, Daniel,

I know you needed to take a break from admin responsibilities and I probably feel like that every week! But it's great to see you back, helping out in AFDLand. You've been missed in our regular rotations of AFD closers. Glad to have you back! Liz 06:33, 26 December 2024 (UTC)

Thanks Liz! Yesterday and today have been reasonably quiet for me so thought I'd jump in and help out. Life gets busy again tomorrow for a few weeks but I'll do whatever I can to help! Cheers, Daniel (talk) 06:42, 26 December 2024 (UTC)