164:. However, on the English Knowledge (XXG), it currently states that "machine translation is disabled for all users and this tool is limited to extended confirmed editors." As a result, only manual translation on the English Knowledge (XXG) is supported by the tool, though some users have used translation to Simple English as a workaround. Relatedly, there is a section of the Help:Translation page with the broad advice: "avoid machine translations." However, this guidance was last edited in 2016, and the state of the art for machine translation has advanced significantly since then, meriting a re-examination of that advice.
42:
104:
is used on a number of
Knowledge (XXG) and Wikimedia projects. This may be directly involved with creation of text content, or in support roles related to evaluating article quality, adding metadata, or generating images. As with any machine-generated content, care must be used when employing AI at
182:
in 2022 has led to increased curiosity in using generative AI to help compose
Knowledge (XXG) articles. The status of machine-generated text from tools such as ChatGPT is generally accepted to be public domain, so the copyright issues are not a blocker to using the generated text from a legal
280:- This project used Met Museum tagging info to train a machine learning system to help predict new "depiction" recommendations for Wikidata. This resulted in a new Wikidata Game that helped add more than 4,000 new depiction (P180) statements to Wikidata. See the Met Museum blog post by
477:
456:(wikitext markup, table rotation, reference analysis, article improvement suggestions, plot summarization, reference- and infobox-based expansion, proseline repair, uncited text tagging, table formatting and color schemes)
270:
Started in 2019, "The computer-aided tagging tool is a feature in development by the
Structured Data on Commons team to assist community members in identifying and labeling depicts statements for Commons files." See:
340:
199:
360:
425:
505:
108:
When exploring AI techniques and systems, the community consensus is to prefer human decisions over machine-generated outcomes until the implications are better understood.
195:) have noted positive aspects of machine generated text, but a serious warning that content must be checked for facts and accuracy and never used straight from ChatGPT.
526:
17:
206:
395:
184:
301:
65:; rather, its purpose is to explain certain aspects of Knowledge (XXG)'s norms, customs, technicalities, or practices. It may reflect differing levels of
272:
501:
521:
326:
160:
used across
Wikimedia projects that can use the output of machine translation from one Knowledge (XXG) article to another, using services like
62:
441:
495:
489:
483:
366:
332:
21:
320:
188:
173:
58:
371:
354:
434:
50:
28:
459:
141:
453:
220:
187:, which advises to make sure content is adjusted for style and that reliable sources are used. Conversations on the
157:
471:
426:"Designing ethically with AI: How Wikimedia can harness machine learning in a responsible and human-centered way"
296:
231:
136:
models that have been trained in order to report article quality or vandalism. This is used in tools such as
101:
409:
285:
185:
Help:Adding open license text to
Knowledge (XXG)#Converting and adding open license text to Knowledge (XXG)
462:(suggestions for article improvement, explanations of unclear maintenance templates based on article text)
70:
66:
465:
225:
381:
207:
Knowledge (XXG):Village pump (policy)/Archive 179#Knowledge (XXG) response to chatbot-generated content
80:
263:
institutions to help supplement image keyword data with machine learning efforts. Among them include:
240:
284:: "Combining AI and Human Judgment to Build Knowledge about Art on a Global Scale," March 4, 2019,
215:
192:
329:, a draft of a proposed policy on using computer-generated content in general on Knowledge (XXG)
323:, an essay on using LLMs (textual generative AI) to produce or modify content on Knowledge (XXG)
260:
125:
386:
236:
161:
153:
133:
137:
335:, a group of editors focusing on the issue of non-policy-compliant LLM-originated content
228:- also: experiments with generating Wikidata Quickstatements from fuzzy date descriptions
508:(humorous April Fools' nominations generated almost entirely by large language models).
515:
375:
349:
281:
144:
to measure the outcomes of classwork, edit-a-thons, or organized editing campaigns.
498:(identifying sourced and unsourced information, only six of seven tests successful)
478:
Knowledge (XXG):Using neural network language models on
Knowledge (XXG)/Transcripts
105:
scale or in applying it where the community consensus is to exercise more caution.
410:"Combining AI and Human Judgment to Build Knowledge about Art on a Global Scale"
390:
article which largely deals with the history and implications of 'Artwork title'
247:
Wikidata game that uses distilBERT and ML, analyzing
Knowledge (XXG) categories.
492:(identifying sourced and unsourced information, including a non-English source)
128:) was started in 2015 as a project of the Wikimedia Foundation, and provides a
307:
361:
m:Research:Implications of ChatGPT for knowledge integrity on
Knowledge (XXG)
244:
205:
A major community discussion took place on
Village Pump (policy) found at:
341:
Knowledge (XXG):Using neural network language models on
Knowledge (XXG)
200:
Knowledge (XXG):Using neural network language models on Knowledge (XXG)
179:
140:
to help immediately revert vandalism, or in evaluation tools like the
116:
AI-related efforts on Knowledge (XXG) include but are not limited to:
468:(PyWikiBot code, writing from scratch, Wikidata parsing, CSV parsing)
27:"Knowledge (XXG):AI" redirects here. For the article incubator, see
357:
developed from raw LLM output (before this page had been developed)
435:"How we're using machine learning to visually enrich Wikidata"
36:
154:
Help:Translation#English Knowledge (XXG) policy requirements
480:(showcases several actual mainspace LLM-assisted copyedits)
198:
A good general page looking at the issues can be found at:
382:
Should ChatGPT Be Used to Write Knowledge (XXG) Articles?
363:, an ongoing (as of July 2023) Wikimedia research project
506:
Knowledge (XXG):Articles for deletion/Planet of the Apes
88:
378:, a surviving article developed from raw LLM output
343:, an essay about large language models specifically
183:standpoint. These issues are generally governed by
18:
Knowledge (XXG):WikiProject Artificial Intelligence
273:c:Commons:Structured data/Computer-aided tagging
486:(identifying sourced and unsourced information)
310:, video essay on AI images and Knowledge (XXG)
502:Knowledge (XXG):Articles for deletion/ChatGPT
396:Artificial intelligence in Wikimedia projects
8:
527:Knowledge (XXG) and artificial intelligence
442:meta:Research:Ethical and human-centered AI
124:The Objective Revision Evaluation Service (
448:Demonstrations of generative AI using LLMs
327:Knowledge (XXG):Computer-generated content
367:m:Wikilegal/Copyright Analysis of ChatGPT
211:Some user experiences can be found here:
297:Wikimedia Commons and AI generated media
63:Knowledge (XXG)'s policies or guidelines
333:Knowledge (XXG):WikiProject AI Cleanup
22:Knowledge (XXG):WikiProject AI Cleanup
321:Knowledge (XXG):Large language models
174:Knowledge (XXG):Large language models
7:
424:Morgan, Jonathan T. (18 July 2019).
302:AI images and German Knowledge (XXG)
496:User:WeatherWriter/LLM Experiment 3
490:User:WeatherWriter/LLM Experiment 2
484:User:WeatherWriter/LLM Experiment 1
278:Metropolitan Museum of Art Tagging
14:
522:Knowledge (XXG) information pages
29:Knowledge (XXG):Article Incubator
191:and in some test articles (i.e.
40:
34:Knowledge (XXG) information page
259:– There have been efforts from
433:Redi, Miriam (14 March 2018).
1:
460:User:JPxG/LLM demonstration 2
412:. Metropolitan Museum of Art.
408:Lih, Andrew (March 4, 2019).
178:The explosion of interest in
142:Program and Events Dashboard
454:User:JPxG/LLM demonstration
221:User:JPxG/LLM demonstration
543:
171:
78:
26:
16:For the WikiProjects, see
15:
472:User:DraconicDark/ChatGPT
232:User:DraconicDark/ChatGPT
152:Guidance can be found at
158:Content Translation Tool
437:. Wikimedia Foundation.
428:. WIkimedia Foundation.
168:Article text generation
102:Artificial intelligence
353:, a surviving article
304:, results of a meeting
268:Computer aided tagging
466:User:Fuzheado/ChatGPT
226:User:Fuzheado/ChatGPT
308:A Battle for Reality
59:encyclopedic article
241:Wikidata:Wwwyzzerdd
252:Images and Commons
216:Talk:Artwork title
99:
98:
534:
474:(lead expansion)
438:
429:
413:
291:Image generation
237:User:BrokenSegue
162:Google Translate
148:Text translation
134:machine learning
120:Revision scoring
91:
51:information page
44:
43:
37:
542:
541:
537:
536:
535:
533:
532:
531:
512:
511:
450:
432:
423:
420:
407:
404:
372:Initial version
317:
254:
176:
170:
150:
122:
114:
95:
94:
87:
83:
75:
74:
41:
35:
32:
25:
12:
11:
5:
540:
538:
530:
529:
524:
514:
513:
510:
509:
499:
493:
487:
481:
475:
469:
463:
457:
449:
446:
445:
444:
439:
430:
419:
416:
415:
414:
403:
400:
399:
398:
393:
392:
391:
369:
364:
358:
345:
344:
337:
336:
330:
324:
316:
313:
312:
311:
305:
299:
288:
287:
275:
257:Image metadata
253:
250:
249:
248:
234:
229:
223:
218:
169:
166:
149:
146:
130:revision score
121:
118:
113:
110:
97:
96:
93:
92:
84:
79:
76:
56:
55:
47:
45:
33:
13:
10:
9:
6:
4:
3:
2:
539:
528:
525:
523:
520:
519:
517:
507:
503:
500:
497:
494:
491:
488:
485:
482:
479:
476:
473:
470:
467:
464:
461:
458:
455:
452:
451:
447:
443:
440:
436:
431:
427:
422:
421:
417:
411:
406:
405:
401:
397:
394:
389:
388:
383:
380:
379:
377:
376:Artwork title
373:
370:
368:
365:
362:
359:
356:
352:
351:
350:Artwork title
347:
346:
342:
339:
338:
334:
331:
328:
325:
322:
319:
318:
314:
309:
306:
303:
300:
298:
295:
294:
293:
292:
286:
283:
279:
276:
274:
269:
266:
265:
264:
262:
258:
251:
246:
242:
238:
235:
233:
230:
227:
224:
222:
219:
217:
214:
213:
212:
209:
208:
203:
201:
196:
194:
193:Artwork title
190:
186:
181:
175:
167:
165:
163:
159:
156:. There is a
155:
147:
145:
143:
139:
135:
131:
127:
119:
117:
111:
109:
106:
103:
90:
86:
85:
82:
77:
72:
68:
64:
61:, nor one of
60:
57:It is not an
54:
52:
46:
39:
38:
30:
23:
19:
385:
348:
290:
289:
277:
267:
256:
255:
210:
204:
197:
189:Village Pump
177:
151:
129:
123:
115:
112:Applications
107:
100:
48:
49:This is an
516:Categories
282:Andrew Lih
172:See also:
138:ClueBot NG
418:Wikimedia
355:initially
67:consensus
315:See also
132:against
81:Shortcut
402:General
245:Psychiq
180:ChatGPT
71:vetting
387:Slate
89:WP:AI
504:and
384:, a
261:GLAM
243:and
126:ORES
69:and
20:and
374:of
518::
239:-
202:.
73:.
53:.
31:.
24:.
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.