This is an old revision of this page, as edited by Herostratus (talk | contribs) at 01:39, 17 July 2021 (→WHOOP WHOOP TERRAIN TERRAIN: nmh). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.
Revision as of 01:39, 17 July 2021 by Herostratus (talk | contribs) (→WHOOP WHOOP TERRAIN TERRAIN: nmh)(diff) ← Previous revision | Latest revision (diff) | Newer revision → (diff)| Skip to table of contents |
False positives and false negativesIf you believe that ClueBot NG has mistakenly identified a good edit as vandalism, please follow the directions in the warning it gave or click here. Please do not report it on this talk page. It takes less time to report the case to the correct location, and we can handle it more effectively there.If you believe that ClueBot NG has missed an edit that is vandalism, again do not report it here. ClueBot is unable to catch all vandalism. Just revert the edit and warn the editor. ClueBot NG Links!Report False Positives • Frequently Asked Questions Purpose of this PageThis page is for comments on or questions about the ClueBots.
The current status of ClueBot NG is: Running
The current status of ClueBot III is: Running
Praise should go on the praise page. Barnstars and other awards should go on the awards page.
Use the "new section" button at the top of this page to add a new section. Use the link above each section to edit that section.
This page is automatically archived by ClueBot III.
The ClueBots' owner or someone else who knows the answer to your question will reply on this page.
| ClueBots | |
|---|---|
| ClueBot NG/Anti-vandalism · ClueBot II/ClueBot Script |  |
| ClueBot III/Archive · Talk page for all ClueBots |
 | Beware! This user's talk page is monitored by talk page watchers. Some of them even talk back. |
He mean
He removed one of my changes 😔 Snow foxy (talk) 19:55, 8 July 2021 (UTC)
Agree Djensje (talk) 03:01, 10 July 2021 (UTC)
Thank you very much!
Thank you for fixing the "List of Pixar Films" page! I was fixing the page at the same time you were and didn't realize it until I reloaded the page. Much thanks indeed! Red4Smash (talk) 18:55, 11 July 2021 (UTC)
WHOOP WHOOP TERRAIN TERRAIN
I've seen (very rarely) the robot remove bad edits that weren't actually vandalism, which makes for a bad edit summary but at least the removed material needed to be removed.
But here the edits were not only not vandalism, but they were actually fine (if not perfect), particularly since by a brand new editor. A promising new editor who was confused as to why she was being called a vandal right off, which is not good.
I hate to say it, but it looks like ClueBot violated the First Law of Robotics here, assuming you define "injury" to include "insult" which I would.
My understanding is that the robot is beyond human control at this point, but here's a suggestion that I think will help: change the edit summaries, from "Reverting possible vandalism" to something milder. Like "Reverting edit per ClueBot algorithm" or something. ("Algorithm", I don't know, pick a better term.) That gets the job done and its simply descriptive so it doesn't insult anyone (I know that this's quite rare, but still). Herostratus (talk) 21:43, 16 July 2021 (UTC)
- @Herostratus: I tried to report a false positive in June, not only could I not report it at the page you're meant to report on (persistent "bad captcha" message), my post here was also ignored. Beyond human control is right. DuncanHill (talk) 22:42, 16 July 2021 (UTC)
- User:DuncanHill, its my understanding, if I have this right, is that it's not a program like we might think of where they can go in and tweak the code. It's an AI, it has taught itself what to do, the creators don't really know why it does what it does, exactly, and they can't really control it. That's my understanding. Herostratus (talk) 01:39, 17 July 2021 (UTC)
- The bot's wording very intentionally never accuses the editor of vandalism. The edit summary says "possible vandalism". The initial warning simply states that a computer program has undone the edit, saying nothing about vandalism. The second warning states that the edits appear to be vandalism to the bot. It's not until warning number 3 that the bot's messaging changes to directly claim vandalism.
- A further note, about that edit in particular: While this is not what the bot is looking for, it's far from perfect. No sources or references, not formatted like the rest of the entries (not italics, a hyphen instead of an ndash), inconsistent spacing and punctuation, "watermelon like" should be "watermelon-like", "that grown" should be "that grows" or "that is grown", "remote" should be "remove", "prosesed" should be "processed", "world famous" should be "world-famous", "liqeur" should be "liqueur", "on ground" should be "on the ground" (or something else?), "the jam is eaten" should be its own sentence or "and is eaten", "kan" should be "can", and so on. The bot is not designed to pick up on any one of these things in particular, but it does analyze most of this, and taken together with common patterns found in other vandalism, it determined this looked enough like vandalism. -- Cobi 23:41, 16 July 2021 (UTC)
- @Cobi: Any comment on the failure of the report false positive mechanism I raised nearly a month ago? DuncanHill (talk) 23:46, 16 July 2021 (UTC)
- Wrong answer Cobi. Are you associated with this device in some way are you just an interested observer?
- @Cobi: Any comment on the failure of the report false positive mechanism I raised nearly a month ago? DuncanHill (talk) 23:46, 16 July 2021 (UTC)
- First of all, no it's not OK for me to say to another editor "I reverted your edit, it's vandalism" if it's not. I can't get off the hook by saying "possible vandalism". That's not going to be considered polite and collegial at ANI either. If I were to say to an editor "you're possibly an asshole" how far am I going to get with that.
- Second of all, colleague, those edits are fine. Particularly for a brand new editor. They are a useful addition to article, I think. Yes they need to be addressed. Tagging for "citation needed" would be OK. Going in and, you know, fixing the formatting stuff would be OK. That is how we do, here. Straight out deleting the contributions is closer to vandalism than the original writing of them. Let me count the ways.
- 1) The material is very likely true.
- 2) It's not really important key material like the date of the Franco-German War or whatever. Nor is it contentious or making some point or something.
- 3) There are probably obtainable refs out there. In fact, I looked, and there are. (That's not counting the refs at the main articles for the subject.)
- 4) I mean it was just written. How about tagging the material and giving people a reasonable to time to put in the refs.
- 5) It was a brand new editor, and a promising one in my opinion. Accusing editors like that of being vandals -- sorry, maybe vandals is not how you grow the project.
- 6) I mean, most or all of the other entries in that section don't have references. They should, yes, so why not address these longer-standing violations than picking on this guy in particular.
- If a human had done this, I would have scolded her, and the "vandalism" thing would have probably warranted a warning. Robots don't get a pass. I get that the robot can't do things like figure out of the existing entries don't have refs and changing its behavior based on that, but so. A human editor like that might be subject to scrutiny on competence grounds. Robots don't get a pass.
- I don't know what you're on about with the levels. South African cuisine: Revision history has one entry by ClueBot. It says "Reverting possible vandalism by " (emphasis added). The editor noticed that and is confused and upset. That matters.
- Yes it is true that on the persons's talk page it is milder. I haven't studied ClueBot's talk page messages, which is why my post addressed, and only addressed, its edit summary.
- I love ClueBot. It's really really helpful, incredibly accurate, and it's an amazing machine too. But, a mistake happened. Nobody is freaking out, we know that no system is perfect. Maybe we can discuss it fruitfully. Maybe something can be done, maybe not (I wouldn't think that changing the edit summaries would be extremely difficult, if the desire is there). Maybe it's just one of those things that we have to live with because overall the device is a net positive (by a lot). Maybe this is evidence of a new trend of edits, maybe not. Maybe other things, I don't know, but I do know when an error occurs, "no it didn't" isn't a functional reply.
- So what's next? What'd be a good venue for a fruitful discussion of whether and how the edit summaries might be adjusted, do you think? Herostratus (talk) 01:39, 17 July 2021 (UTC)