402:
130:
367:
383:
496:, claiming that these companies have infringed the rights of millions of artists by training AI tools on five billion images scraped from the web without the consent of the original artists. The plaintiffs' complaint has been criticized for technical inaccuracies, such as incorrectly claiming that "a trained diffusion model can produce a copy of any of its Training Images", and describing Stable Diffusion as "merely a complex collage tool". In addition to copyright infringement, the plaintiffs allege unlawful competition and violation of their
79:
349:" (TDM) in their copyright statutes including the United Kingdom, Germany, Japan, and the EU. Unlike the EU, the United Kingdom prohibits data mining for commercial purposes but has proposed this should be changed to support the development of AI: "For text and data mining, we plan to introduce a new copyright and database exception which allows TDM for any purpose. Rights holders will still have safeguards to protect their content, including a requirement for lawful access." As of June 2023, a clause in the draft EU
444:(including his particular brush strokes, use of colour, perspective, and so on), and a user can engineer a prompt such as "an astronaut riding a horse, by Picasso" to cause the model to generate a novel image applying the artist's style to an arbitrary subject. However, an artist's overall style is generally not subject to copyright protection.
426:-series models) as variously over 1% for exact duplicates or up to about 7%. This is potentially a security risk and a copyright risk, for both users and providers. As of August 2023, major consumer LLMs have attempted to mitigate these problems, but researchers have still been able to prompt leakage of copyrighted material.
544:
against Meta and OpenAI, arguing that in addition to copyright infringement for training their engines on their works, that products produced from the AI engines were derivative works and also copyright infringements. The two suits against OpenAI were combined (during which Awad left the suit) and by
198:
for the
Federal Circuit upheld this decision. In the subsequent rule-making, the USPTO allows for human inventors to incorporate the output of artificial intelligence, as long as this method is appropriately documented in the patent application. However, it may become virtually impossible as when the
106:
has interpreted this as being limited to works "created by a human being", declining to grant copyright to works generated without human intervention. Some have suggested that certain AI generations might be copyrightable in the U.S. and similar jurisdictions if it can be shown that the human who ran
531:
In July 2023, authors Paul
Tremblay and Mona Awad filed a lawsuit against OpenAI, alleging that ChatGPT, OpenAI's language model, used their copyrighted books without permission. ChatGPT's accurate summaries of their works suggest unauthorized content use. Two separate lawsuits were filed by authors
169:
One such recent development is the use of sophisticated artificial intelligence ("AI") technologies capable of producing expressive material. These technologies "train" on vast quantities of preexisting human-authored works and use inferences from that training to generate new content. Some systems
265:
As noted by a lawyer and AI art creator, the challenge for intellectual property regulators, legislators and the courts is how to protect human creativity in a technologically neutral fashion whilst considering the risks of automated AI factories. AI tools have the ability to autonomously create a
235:
However, the computer generated work law under UK law relates to autonomous creations by computer programs. Individuals using AI tools will usually be the authors of the works assuming they meet the minimum requirements for copyright work. The language used for computer generated work relates, in
110:
Proponents of this view suggest that an AI model may be viewed as merely a tool (akin to a pen or a camera) used by its human operator to express their creative vision. For example, proponents argue that if the standard of originality can be satisfied by an artist clicking the shutter button on a
508:
inclined to dismiss most of the lawsuit filed by
Andersen, McKernan, and Ortiz but allowed them to file a new complaint. Judge Orrick later dismissed all but one claim, that of copyright infringement towards Stability AI, in October 2023. However, after refiling on some of the eliminated claims,
173:
The resulting output may be textual, visual, or audio, and is determined by the AI based on its design and the material it has been trained on. These technologies, often described as "generative AI," raise questions about whether the material they produce is protected by copyright, whether works
240:
In so far as each composite frame is a computer generated work then the arrangements necessary for the creation of the work were undertaken by Mr Jones because he devised the appearance of the various elements of the game and the rules and logic by which each frame is generated and he wrote the
156:
The Board finds that the Work contains more than a de minimis amount of content generated by artificial intelligence ("AI"), and this content must therefore be disclaimed in an application for registration. Because Mr. Allen is unwilling to disclaim the AI-generated material, the Work cannot be
282:
the AI can analyze. A deep learning model identifies patterns linking the encoded text and image data and learns which text concepts correspond to elements in images. Through repetitive testing, the model refines its accuracy by matching images to text descriptions. The trained model undergoes
245:
The UK government has consulted on the use of generative tools and AI in respect of intellectual property leading to a proposed specialist Code of
Practice: "to provide guidance to support AI firms to access copyrighted work as an input to their models, whilst ensuring there are protections on
161:
As AI is increasingly used to generate literature, music, and other forms of art, the U.S. Copyright Office has released new guidance emphasizing whether works, including materials generated by artificial intelligence, exhibit a 'mechanical reproduction' nature or are the 'manifestation of the
337:
that datasets are so large that "there is no plausible option simply to license all . So allowing copyright claim is tantamount to saying, not that copyright owners will get paid, but that the use won't be permitted at all." Other scholars disagree; some predict a similar outcome to the U.S.
111:
camera, then perhaps artists using generative AI should get similar deference, especially if they go through multiple rounds of revision to refine their prompts to the AI. Other proponents argue that the
Copyright Office is not taking a technology neutral approach to the use of AI or
1383:
Hans, Abhimanyu; Wen, Yuxin; Jain, Neel; Kirchenbauer, John; Kazemi, Hamid; Singhania, Prajwal; Singh, Siddharth; Somepalli, Gowthami; Geiping, Jonas; Bhatele, Abhinav; Goldstein, Tom (June 14, 2024). "Be like a
Goldfish, Don't Memorize! Mitigating Memorization in Generative LLMs".
519:
Getty filed another suit against
Stability AI in a U.S. district court in Delaware in February 2023. The suit again alleges copyright infringement for the use of Getty's images in the training of Stable Diffusion, and further argues that the model infringes Getty's
429:
Under U.S. law, to prove that an AI output infringes a copyright, a plaintiff must show the copyrighted work was "actually copied", meaning that the AI generates output which is "substantially similar" to their work, and that the AI had access to their work.
322:
succeed, this may shift the balance of power in favour of large corporations such as Google, Microsoft, and Meta which can afford to license large amounts of training data from copyright holders and leverage their proprietary datasets of user-generated data.
469:, an AI-powered code editing tool trained on public GitHub repositories, violated the copyright of the repositories' authors, noting that the tool was able to generate source code which matched its training data verbatim, without providing attribution.
1425:
Liu, Yang; Yao, Yuanshun; Ton, Jean-Francois; Zhang, Xiaoying; Guo, Ruocheng; Cheng, Hao; Klochkov, Yegor; Taufiq, Muhammad Faaiz; Li, Hang (August 10, 2023). "Trustworthy LLMs: a Survey and
Guideline for Evaluating Large Language Models' Alignment".
1191:
956:
166:...because the Office receives roughly half a million applications for registration each year, it sees new trends in registration activity that may require modifying or expanding the information required to be disclosed on an application.
241:
relevant computer program. In these circumstances I am satisfied that Mr Jones is the person by whom the arrangements necessary for the creation of the works were undertaken and therefore is deemed to be the author by virtue of s.9(3)
394:
Generative AI models may produce outputs that are virtually identical to images from their training set. The research paper from which this example was taken was able to produce similar replications for only 0.03% of training
580:
sued
Microsoft and OpenAI in April 2024 over copyright infringement related to the use of their news articles for training data, as well as for output that creates false and misleading statements that are attributed to the
421:
is the emergent phenomenon of LLMs to repeat long strings of training data, and it is no longer related to overfitting. Evaluations of controlled LLM output measure the amount memorized from training data (focused on
596:, AI models that can take text input to create songs with both lyrics and backing music, in separate lawsuits in June 2024, alledging that both AI models were trained without consent with music from the labels.
66:'s exclusive right to control reproduction, unless covered by exceptions in relevant copyright laws. Additionally, using a model's outputs might violate copyright, and the model creator could be accused of
287:
involves making copies of copyrighted works, this has raised the question of whether this process infringes the copyright holder's exclusive right to make reproductions of their works, or if it falls use
230:
In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are
433:
In the course of learning to statistically model the data on which they are trained, deep generative AI models may learn to imitate the distinct style of particular authors in the training set. Since
844:
90:
Since most legal jurisdictions only grant copyright to original works of authorship by human authors, the definition of "originality" is central to the copyright status of AI-generated works.
266:
range of material that is potentially subject to copyright (music, blogs, poetry, images, and technical papers) or other intellectual property rights (patents and design rights).
1700:
174:
consisting of both human-authored and AI-generated material may be registered, and what information should be provided to the Office by applicants seeking to register them.
1448:
409:
using the prompt "an astronaut riding a horse, by
Picasso". Generative image models are adept at imitating the visual style of particular artists in their training set.
549:
of the Northern District of California threw out all but one claim related to the use of the author's copyrighted works as part of the training data for the AI model.
283:
validation to evaluate its skill in generating or manipulating new images using only the text prompts provided after the training process. Because assembling these
236:
respect of AI, to the ability of the human programmers to have copyright in the autonomous productions of the AI tools (i.e. where there is no direct human input):
1488:
51:. As of 2023, there are several pending U.S. lawsuits challenging the use of copyrighted data to train AI models, with defendants arguing that this falls under
1472:
884:
Valinasab, Omid, "Big Data Analytics to Automate Patent Disclosure of Artificial Intelligence’s Inventions." U.S.F. Intell. Prop. & Tech. L.J. 133 (2023).
162:
author's own creative conception'. The U.S. Copyright Office published a Rule in March 2023 on a range of issues related to the use of AI, where they stated:
1641:
1533:
569:
further claimed that fair use claims made by these AI companies were invalid since the generated information around news stories directly competes with the
509:
Orrick agreed in August 2024 to include some of these additional claims against the AI companies, which included both copyright and trademark infringements.
413:
In some cases, deep learning models may replicate items in their training set when generating output. This behaviour is generally considered an undesired
1561:
1257:
1131:
1105:
1289:
Well-constructed AI systems generally do not regenerate, in any nontrivial portion, unaltered data from any particular work in their training corpus
1782:
437:
in the U.S. and other jurisdictions, an AI may also produce infringing content in the form of novel works which incorporate fictional characters.
1507:
122:
or limited human creativity. For works using AI tools, the Copyright Office has made the test a different one i.e. whether there is no more than
1756:
1407:
214:
before releasing new generative AI systems. These companies would have to file these documents 30 days before publicly showing their AI tools.
62:
from the Internet, often utilizing copyrighted material. When assembling training data, the sourcing of copyrighted works may infringe on the
1728:
1345:
434:
1846:
278:
large data sets from the Internet such as publicly available images and the text of web pages. The text and images are then converted into
309:, which was ultimately found to be fair use, because the scanned content was not made publicly available, and the use was non-expressive.
1674:
981:
857:
1079:
721:
146:
This difference in approach can be seen in the recent decision in respect of a registration claim by Jason Matthew Allen for his work
585:
179:
295:
U.S. machine learning developers have traditionally believed this to be allowable under fair use because using copyrighted work is
1810:
1408:"ByteDance AI researchers say OpenAI now tries to hide that ChatGPT was trained on J.K. Rowling's copyrighted Harry Potter books"
223:
908:
664:
639:
207:
894:
1237:
353:
would require generative AI to "make available summaries of the copyrighted material that was used to train their systems".
222:
Other jurisdictions include explicit statutory language related to computer-generated works, including the United Kingdom's
1473:
James Vincent "AI art tools Stable Diffusion and Midjourney targeted with copyright lawsuit" The Verge, 16 January, 2023.
305:
1281:
557:
has sued Microsoft and OpenAI in December 2023, claiming that their engines were trained on wholesale articles from the
28:
418:
148:
134:
1862:
1589:
944:
440:
A generative image model such as Stable Diffusion is able to model the stylistic characteristics of an artist like
1659:
1282:
Comment Regarding Request for Comments on Intellectual Property Protection for Artificial Intelligence Innovation
199:
inner workings and the use of AI in inventive transactions are not adequately understood or are largely unknown.
83:
614:
1192:"Artificial Intelligence and Intellectual Property: copyright and patents: Government response to consultation"
246:
generated output to support right holders of copyrighted work". The U.S. Copyright Office recently published a
107:
the AI program exercised sufficient originality in selecting the inputs to the AI or editing the AI's output.
957:
Artificial Intelligence and Intellectual Property: copyright and patents: Government response to consultation
247:
115:
tools. For other creative expressions (music, photography, writing) the test is effectively whether there is
1218:"Heuristic Analysis for Security, Privacy and Bias of Text Generative AI: GhatGPT-3.5 case as of June 2023"
762:"Popular A.I. services for creating images are legal minefields for artists seeking payment for their work"
546:
1614:
1044:
Henderson, Peter; Li, Xuechen; Jurafsky, Dan; Hashimoto, Tatsunori; Lemley, Mark A.; Liang, Percy (2023).
417:
of a model by AI developers, and has in previous generations of AI been considered a manageable problem.
1840:
968:
186:
of patents credits solely to AI authors in February 2024, following an August 2023 ruling in the case
1000:
401:
1314:
790:
743:
190:. In this case, the Patent Office denied grant to patents created by Stephen Thaler's AI program,
932:
210:
in April 2024. If passed, the bill would require AI companies to submit copyrighted works to the
129:
1867:
845:
Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence
761:
481:
262:
Internet Court issued a decision recognizing copyright in AI-generated images in a litigation.
211:
82:
The United States Copyright Office has declared that works not created by a human author, like
32:
324:
195:
103:
99:
703:
44:
1315:"Near-Duplicate Sequence Search at Scale for Large Language Model Memorization Evaluation"
8:
67:
36:
1538:
1427:
1385:
1337:
577:
553:
537:
505:
497:
296:
1217:
1341:
1233:
1229:
1173:
345:
Several jurisdictions have explicitly incorporated exceptions allowing for "text and
1590:"Getty Images suing the makers of popular AI art tool for allegedly stealing photos"
806:
Second Request for Reconsideration for Refusal to Register Théâtre D'opéra Spatial (
382:
1760:
1534:"Artists Lose First Round of Copyright Infringement Case Against AI Art Generators"
1329:
1225:
1165:
406:
373:
333:
284:
112:
78:
63:
40:
1025:
Murray, Michael (2023). "Generative AI Art: Copyright Infringement and Fair Use".
366:
1287:(Report). United States Patent and Trademark Office. p. 9. PTO–C–2019–0038.
533:
339:
500:
in relation to AI tools' ability to create works in the style of the plaintiffs
1642:"Getty sues Stability AI for copying 12M photos and imitating famous watermark"
541:
477:
473:
466:
1856:
1177:
441:
183:
152:
created using Midjourney and an upscaling tool. The Copyright Office stated:
24:
513:
485:
314:
300:
275:
59:
1157:
1080:"The scary truth about AI copyright is nobody knows what will happen next"
1045:
722:"The scary truth about AI copyright is nobody knows what will happen next"
388:
An image generated by Stable Diffusion using the prompt "Anne Graham Lotz"
1675:"OpenAI faces copyright lawsuit from authors Mona Awad and Paul Tremblay"
1169:
516:
for using its images in their training data without purchasing a license.
414:
346:
328:
279:
203:
170:
operate in response to a user's textual instruction, called a "prompt."
1489:"Artists file class-action lawsuit against AI image generator companies"
1258:"Stable Diffusion copyright lawsuits could be a legal earthquake for AI"
1132:"Stable Diffusion copyright lawsuits could be a legal earthquake for AI"
1106:"Stable Diffusion copyright lawsuits could be a legal earthquake for AI"
820:
493:
489:
139:
117:
1562:"Artists' lawsuit against Stability AI and Midjourney gets more punch"
194:
due to the lack of a "natural person" on the patents' authorship. The
1815:
1733:
1705:
1701:"Sarah Silverman is suing OpenAI and Meta for copyright infringement"
1566:
1432:
525:
521:
454:
319:
1333:
982:"Beijing Internet Court Recognizes Copyright in AI-Generated Images"
969:
https://www.govinfo.gov/content/pkg/FR-2023-08-30/pdf/2023-18624.pdf
909:"New bill would force AI companies to reveal use of copyrighted art"
1847:
Getting the Innovation Ecosystem Ready for AI: An IP policy toolkit
1783:"Eight newspapers sue OpenAI, Microsoft for copyright infringement"
1390:
289:
250:
and request for comments following its 2023 Registration Guidance.
52:
58:
Popular deep learning models are trained on mass amounts of media
615:"Artificial Intelligence Copyright Challenges in US Courts Surge"
589:
259:
48:
20:
35:
occurs when the generative AI is trained or used. This includes
1757:"New York Times and authors on 'fair use' of copyrighted works"
1508:"US judge finds flaws in artists' lawsuit against AI companies"
462:
458:
350:
1729:"Sarah Silverman's lawsuit against OpenAI partially dismissed"
1275:
588:(RIAA) and several major music labels sued the developers of
423:
191:
1043:
807:
593:
1449:"The lawsuit that could rewrite the rules of AI copyright"
945:
The UK government's code of practice on copyright and AI.
1787:
1313:
Peng, Zhencan; Wang, Zhizhi; Deng, Dong (June 13, 2023).
73:
1811:"Major record labels sue AI company behind 'BBL Drizzy'"
858:"'Thaler v. Perlmutter': AI Output is Not Copyrightable"
1382:
1300:
1222:
2023 IEEE International Conference on Computing (ICOCO)
16:
How copyright law applies to the training and use of AI
704:"Generative Artificial Intelligence and Copyright Law"
70:
and held responsible for that copyright infringement.
665:"What is the Copyright Status of AI Generated Works?"
573:
and impacts the newspaper's commercial opportunities.
512:
In January 2023, Stability AI was sued in London by
435:
fictional characters enjoy some copyright protection
269:
640:"Primer: Training AI Models with Copyrighted Work"
356:
299:, and limited. The situation has been compared to
1660:"The copyright battles against OpenAI have begun"
1001:"Model validation techniques in machine learning"
1854:
1274:See for example OpenAI's comment in the year of
808:SR # 1-11743923581; Correspondence ID: 1-5T5320R
1841:Getty Images (U.S.), Inc. v. Stability AI, Inc.
565:considers infringement of their copyright. The
182:(USPTO) similarly codified restrictions on the
1424:
979:
1440:
1251:
1249:
702:Zirpoli, Christopher T. (February 24, 2023).
453:A November 2022 class action lawsuit against
102:protects "original works of authorship". The
1607:
1370:
1322:Proceedings of the ACM on Management of Data
1312:
933:Nova Production v MazoomaGames EWHC 24 (Ch)
86:, are not eligible for copyright protection.
1754:
1155:
376:included in Stable Diffusion's training set
31:models are raising questions about whether
1246:
1073:
1071:
785:
783:
1633:
1581:
1559:
1431:
1389:
1215:
950:
895:"USPTO says AI models can't hold patents"
821:"Federal Register :: Request Access"
586:Recording Industry Association of America
1639:
1505:
1376:
741:
576:Eight U.S. national newspapers owned by
400:
128:
84:this "selfie" portrait taken by a monkey
77:
1486:
1482:
1480:
1446:
1077:
1068:
780:
744:"Artificial intelligence and copyright"
719:
715:
713:
701:
480:, and Karla Ortiz—filed a class action
224:Copyright, Designs and Patents Act 1988
1855:
1466:
1279:
1156:Lemley, Mark A.; Casey, Bryan (2020).
1024:
998:
737:
735:
208:Generative AI Copyright Disclosure Act
74:Copyright status of AI-generated works
1780:
1726:
1698:
1640:Belanger, Ashley (February 6, 2023).
1216:Mitsunaga, Takuho (October 9, 2023).
1027:SMU Science and Technology Law Review
880:
878:
697:
695:
693:
691:
689:
687:
685:
1808:
1672:
1587:
1477:
1405:
1351:from the original on August 27, 2024
1078:Vincent, James (November 15, 2022).
1050:Journal of Machine Learning Research
1005:LeewayHertz - AI Development Company
980:Aaron Wininger (November 29, 2023).
720:Vincent, James (November 15, 2022).
710:
504:. In July 2023, U.S. District Judge
303:'s scanning of copyrighted books in
1727:David, Emilla (February 13, 2024).
1588:Korn, Jennifer (January 17, 2023).
1531:
1447:Vincent, James (November 8, 2022).
1255:
1129:
1103:
732:
306:Authors Guild, Inc. v. Google, Inc.
13:
1755:O'Brien, Matt (January 10, 2024).
1560:Robertson, Adi (August 13, 2024).
1487:Edwards, Benj (January 16, 2023).
999:Takyar, Akash (November 7, 2023).
875:
682:
524:by generating images with Getty's
29:generative artificial intelligence
14:
1879:
1834:
1532:Cho, Winston (October 30, 2023).
1506:Brittain, Blake (July 19, 2023).
1256:Lee, Timothy B. (April 3, 2023).
1130:Lee, Timothy B. (April 3, 2023).
1104:Lee, Timothy B. (April 3, 2023).
754:
742:Guadamuz, Andres (October 2017).
706:. Congressional Research Service.
270:Training AI with copyrighted data
217:
1230:10.1109/icoco59262.2023.10397858
1046:"Foundation Models and Fair Use"
791:Copyright, AI And Generative Art
381:
365:
180:U.S. Patent and Trademark Office
93:
1802:
1781:Allyn, Bobby (April 30, 2024).
1774:
1748:
1720:
1692:
1666:
1652:
1553:
1525:
1499:
1399:
1364:
1306:
1294:
1268:
1209:
1184:
1149:
1123:
1097:
1037:
1018:
992:
973:
962:
938:
926:
901:
887:
850:
838:
813:
357:Copyright infringing AI outputs
1406:Hays, Kali (August 15, 2023).
800:
657:
632:
607:
472:In January 2023 three artists—
1:
600:
447:
1673:Kris, Jimmy (July 6, 2023).
142:, prompted by Jason M. Allen
7:
1809:Sato, Mia (June 24, 2024).
1699:Davis, Wes (July 9, 2023).
847:, US Copyright Office 2023.
825:unblock.federalregister.gov
126:technological involvement.
10:
1884:
1371:Peng, Wang & Deng 2023
1224:. IEEE. pp. 301–305.
258:On November 27, 2023, the
1619:newsroom.gettyimages.com/
327:scholars Bryan Casey and
1615:"Getty Images Statement"
253:
157:registered as submitted.
1863:Artificial intelligence
1162:SSRN Electronic Journal
986:The National Law Review
547:Araceli MartĂnez-OlguĂn
149:Théâtre D'opéra Spatial
135:Théâtre D'opéra Spatial
1361:Citing Lee et al 2022.
482:copyright infringement
410:
405:An image generated by
243:
233:
212:Register of Copyrights
176:
159:
143:
138:, 2022, created using
87:
33:copyright infringement
959:. UK Government 2023.
545:February 2024, Judge
404:
318:, argues that if the
274:Deep learning models
238:
228:
196:U.S. Court of Appeals
164:
154:
132:
104:U.S. Copyright Office
81:
45:large language models
1170:10.2139/ssrn.3528447
897:. February 14, 2024.
862:New York Law Journal
619:www.natlawreview.com
188:Thaler v. Perlmutter
37:text-to-image models
947:UK Government 2023.
913:amp.theguardian.com
789:Peter Pink-Howitt,
312:Timothy B. Lee, in
68:vicarious liability
1621:. January 17, 2023
1539:Hollywood Reporter
578:Tribune Publishing
554:The New York Times
538:Christopher Golden
498:right of publicity
411:
144:
88:
19:In the 2020s, the
1239:979-8-3503-0268-4
285:training datasets
248:notice of inquiry
98:In the U.S., the
21:rapid advancement
1875:
1828:
1827:
1825:
1823:
1806:
1800:
1799:
1797:
1795:
1778:
1772:
1771:
1769:
1767:
1761:Associated Press
1752:
1746:
1745:
1743:
1741:
1724:
1718:
1717:
1715:
1713:
1696:
1690:
1689:
1687:
1685:
1670:
1664:
1663:
1656:
1650:
1649:
1637:
1631:
1630:
1628:
1626:
1611:
1605:
1604:
1602:
1600:
1585:
1579:
1578:
1576:
1574:
1557:
1551:
1550:
1548:
1546:
1529:
1523:
1522:
1520:
1518:
1503:
1497:
1496:
1484:
1475:
1470:
1464:
1463:
1461:
1459:
1444:
1438:
1437:
1435:
1422:
1420:
1418:
1412:Business Insider
1403:
1397:
1395:
1393:
1380:
1374:
1368:
1362:
1360:
1358:
1356:
1350:
1319:
1310:
1304:
1301:Hans et al. 2024
1298:
1292:
1291:
1286:
1272:
1266:
1265:
1253:
1244:
1243:
1213:
1207:
1206:
1204:
1202:
1188:
1182:
1181:
1153:
1147:
1146:
1144:
1142:
1127:
1121:
1120:
1118:
1116:
1101:
1095:
1094:
1092:
1090:
1075:
1066:
1065:
1063:
1061:
1041:
1035:
1034:
1022:
1016:
1015:
1013:
1011:
996:
990:
989:
977:
971:
966:
960:
954:
948:
942:
936:
930:
924:
923:
921:
919:
905:
899:
898:
891:
885:
882:
873:
872:
870:
868:
854:
848:
842:
836:
835:
833:
831:
817:
811:
804:
798:
787:
778:
777:
775:
773:
758:
752:
751:
739:
730:
729:
717:
708:
707:
699:
680:
679:
677:
675:
669:www.linkedin.com
661:
655:
654:
652:
650:
636:
630:
629:
627:
625:
611:
484:lawsuit against
407:Stable Diffusion
385:
374:Anne Graham Lotz
372:A photograph of
369:
334:Texas Law Review
226:, which states:
64:copyright holder
41:Stable Diffusion
1883:
1882:
1878:
1877:
1876:
1874:
1873:
1872:
1853:
1852:
1837:
1832:
1831:
1821:
1819:
1807:
1803:
1793:
1791:
1779:
1775:
1765:
1763:
1753:
1749:
1739:
1737:
1725:
1721:
1711:
1709:
1697:
1693:
1683:
1681:
1671:
1667:
1662:. July 6, 2023.
1658:
1657:
1653:
1638:
1634:
1624:
1622:
1613:
1612:
1608:
1598:
1596:
1586:
1582:
1572:
1570:
1558:
1554:
1544:
1542:
1530:
1526:
1516:
1514:
1504:
1500:
1485:
1478:
1471:
1467:
1457:
1455:
1445:
1441:
1416:
1414:
1404:
1400:
1381:
1377:
1369:
1365:
1354:
1352:
1348:
1334:10.1145/3589324
1317:
1311:
1307:
1299:
1295:
1284:
1280:OpenAI (2019).
1273:
1269:
1254:
1247:
1240:
1214:
1210:
1200:
1198:
1190:
1189:
1185:
1158:"Fair Learning"
1154:
1150:
1140:
1138:
1128:
1124:
1114:
1112:
1102:
1098:
1088:
1086:
1076:
1069:
1059:
1057:
1042:
1038:
1023:
1019:
1009:
1007:
997:
993:
978:
974:
967:
963:
955:
951:
943:
939:
931:
927:
917:
915:
907:
906:
902:
893:
892:
888:
883:
876:
866:
864:
856:
855:
851:
843:
839:
829:
827:
819:
818:
814:
805:
801:
788:
781:
771:
769:
760:
759:
755:
740:
733:
718:
711:
700:
683:
673:
671:
663:
662:
658:
648:
646:
638:
637:
633:
623:
621:
613:
612:
608:
603:
534:Sarah Silverman
450:
399:
398:
397:
396:
391:
390:
389:
386:
378:
377:
370:
359:
340:music licensing
280:numeric formats
272:
256:
220:
202:Representative
96:
76:
17:
12:
11:
5:
1881:
1871:
1870:
1865:
1851:
1850:
1844:
1836:
1835:External links
1833:
1830:
1829:
1801:
1773:
1747:
1719:
1691:
1665:
1651:
1632:
1606:
1580:
1552:
1524:
1498:
1476:
1465:
1439:
1398:
1375:
1363:
1305:
1293:
1267:
1245:
1238:
1208:
1183:
1148:
1122:
1096:
1067:
1036:
1017:
991:
972:
961:
949:
937:
925:
900:
886:
874:
849:
837:
812:
799:
779:
753:
731:
709:
681:
656:
631:
605:
604:
602:
599:
598:
597:
582:
574:
550:
542:Richard Kadrey
529:
517:
510:
506:William Orrick
478:Kelly McKernan
474:Sarah Andersen
470:
467:GitHub Copilot
449:
446:
393:
392:
387:
380:
379:
371:
364:
363:
362:
361:
360:
358:
355:
297:transformative
271:
268:
255:
252:
219:
218:United Kingdom
216:
95:
92:
75:
72:
15:
9:
6:
4:
3:
2:
1880:
1869:
1868:Copyright law
1866:
1864:
1861:
1860:
1858:
1848:
1845:
1842:
1839:
1838:
1818:
1817:
1812:
1805:
1790:
1789:
1784:
1777:
1762:
1758:
1751:
1736:
1735:
1730:
1723:
1708:
1707:
1702:
1695:
1680:
1676:
1669:
1661:
1655:
1647:
1643:
1636:
1620:
1616:
1610:
1595:
1591:
1584:
1569:
1568:
1563:
1556:
1541:
1540:
1535:
1528:
1513:
1509:
1502:
1494:
1490:
1483:
1481:
1474:
1469:
1454:
1450:
1443:
1434:
1429:
1417:September 15,
1413:
1409:
1402:
1392:
1387:
1379:
1372:
1367:
1347:
1343:
1339:
1335:
1331:
1327:
1323:
1316:
1309:
1302:
1297:
1290:
1283:
1277:
1271:
1263:
1259:
1252:
1250:
1241:
1235:
1231:
1227:
1223:
1219:
1212:
1197:
1193:
1187:
1179:
1175:
1171:
1167:
1163:
1159:
1152:
1137:
1133:
1126:
1111:
1107:
1100:
1085:
1081:
1074:
1072:
1060:September 14,
1055:
1051:
1047:
1040:
1032:
1028:
1021:
1006:
1002:
995:
987:
983:
976:
970:
965:
958:
953:
946:
941:
934:
929:
914:
910:
904:
896:
890:
881:
879:
863:
859:
853:
846:
841:
826:
822:
816:
809:
803:
796:
792:
786:
784:
767:
763:
757:
749:
748:WIPO Magazine
745:
738:
736:
727:
723:
716:
714:
705:
698:
696:
694:
692:
690:
688:
686:
670:
666:
660:
645:
641:
635:
620:
616:
610:
606:
595:
591:
587:
583:
579:
575:
572:
568:
564:
560:
556:
555:
551:
548:
543:
539:
535:
530:
527:
523:
518:
515:
511:
507:
503:
499:
495:
491:
487:
483:
479:
475:
471:
468:
465:alleged that
464:
460:
456:
452:
451:
445:
443:
442:Pablo Picasso
438:
436:
431:
427:
425:
420:
416:
408:
403:
384:
375:
368:
354:
352:
348:
343:
341:
336:
335:
331:argue in the
330:
326:
321:
317:
316:
310:
308:
307:
302:
298:
293:
291:
286:
281:
277:
267:
263:
261:
251:
249:
242:
237:
232:
227:
225:
215:
213:
209:
206:proposed the
205:
200:
197:
193:
189:
185:
184:patentability
181:
175:
171:
167:
163:
158:
153:
151:
150:
141:
137:
136:
131:
127:
125:
121:
119:
114:
108:
105:
101:
100:Copyright Act
94:United States
91:
85:
80:
71:
69:
65:
61:
56:
54:
50:
46:
42:
38:
34:
30:
26:
25:deep learning
22:
1820:. Retrieved
1814:
1804:
1792:. Retrieved
1786:
1776:
1764:. Retrieved
1750:
1738:. Retrieved
1732:
1722:
1710:. Retrieved
1704:
1694:
1682:. Retrieved
1678:
1668:
1654:
1646:Ars Technica
1645:
1635:
1623:. Retrieved
1618:
1609:
1597:. Retrieved
1593:
1583:
1571:. Retrieved
1565:
1555:
1543:. Retrieved
1537:
1527:
1515:. Retrieved
1511:
1501:
1493:Ars Technica
1492:
1468:
1456:. Retrieved
1452:
1442:
1433:2308.05374v2
1415:. Retrieved
1411:
1401:
1378:
1373:, p. 8.
1366:
1353:. Retrieved
1325:
1321:
1308:
1296:
1288:
1278:'s release:
1270:
1262:Ars Technica
1261:
1221:
1211:
1199:. Retrieved
1195:
1186:
1161:
1151:
1139:. Retrieved
1136:Ars Technica
1135:
1125:
1113:. Retrieved
1110:Ars Technica
1109:
1099:
1087:. Retrieved
1083:
1058:. Retrieved
1053:
1049:
1039:
1030:
1026:
1020:
1008:. Retrieved
1004:
994:
985:
975:
964:
952:
940:
928:
916:. Retrieved
912:
903:
889:
865:. Retrieved
861:
852:
840:
828:. Retrieved
824:
815:
802:
794:
770:. Retrieved
765:
756:
747:
725:
672:. Retrieved
668:
659:
647:. Retrieved
643:
634:
622:. Retrieved
618:
609:
570:
566:
562:
561:, which the
558:
552:
514:Getty Images
501:
486:Stability AI
439:
432:
428:
419:Memorization
412:
344:
342:procedures.
332:
315:Ars Technica
313:
311:
304:
301:Google Books
294:
292:allowances.
273:
264:
257:
244:
239:
234:
229:
221:
201:
187:
177:
172:
168:
165:
160:
155:
147:
145:
133:
123:
116:
109:
97:
89:
57:
18:
1849:– WIPO 2024
1625:January 24,
1599:January 22,
1458:December 7,
1355:January 20,
1328:(2): 1–18.
1056:(400): 1–79
867:December 1,
581:newspapers.
415:overfitting
347:data mining
329:Mark Lemley
231:undertaken.
204:Adam Schiff
113:algorithmic
1857:Categories
1573:August 13,
1391:2406.10209
601:References
494:DeviantArt
490:Midjourney
448:Litigation
320:plaintiffs
140:Midjourney
124:de minimis
118:de minimis
1816:The Verge
1794:April 30,
1766:April 30,
1740:April 30,
1734:The Verge
1712:April 30,
1706:The Verge
1567:The Verge
1545:April 30,
1517:August 6,
1453:The Verge
1342:259213212
1201:March 20,
1178:1556-5068
1141:March 20,
1115:March 20,
1089:March 20,
1084:The Verge
1010:March 20,
918:April 13,
830:March 20,
726:The Verge
674:March 19,
649:March 19,
624:March 19,
526:watermark
522:trademark
455:Microsoft
1822:June 24,
1684:July 10,
1346:Archived
810:, 2023).
795:Ramparts
772:June 21,
502:en masse
290:fair use
53:fair use
47:such as
39:such as
1843:filings
1679:DailyAi
1512:Reuters
1423:citing
797:, 2023.
766:Fortune
590:Suno AI
395:images.
260:Beijing
60:scraped
49:ChatGPT
27:-based
1340:
1303:, §2.3
1236:
1196:GOV.UK
1176:
768:. 2023
492:, and
463:OpenAI
459:GitHub
351:AI Act
276:source
1428:arXiv
1386:arXiv
1349:(PDF)
1338:S2CID
1318:(PDF)
1285:(PDF)
1276:GPT-2
571:Times
567:Times
563:Times
559:Times
424:GPT-2
254:China
192:DABUS
1824:2024
1796:2024
1768:2024
1742:2024
1714:2024
1686:2023
1627:2023
1601:2023
1575:2024
1547:2024
1519:2023
1460:2022
1419:2024
1357:2024
1234:ISBN
1203:2024
1174:ISSN
1143:2024
1117:2024
1091:2024
1062:2024
1033:(2).
1012:2024
920:2024
869:2023
832:2024
774:2023
676:2024
651:2024
626:2024
594:Udio
592:and
584:The
540:and
461:and
178:The
43:and
1788:NPR
1594:CNN
1396:§1.
1330:doi
1226:doi
1166:doi
793:,
644:AAF
23:of
1859::
1813:.
1785:.
1759:.
1731:.
1703:.
1677:.
1644:.
1617:.
1592:.
1564:.
1536:.
1510:.
1491:.
1479:^
1451:.
1410:.
1344:.
1336:.
1324:.
1320:.
1260:.
1248:^
1232:.
1220:.
1194:.
1172:.
1164:.
1160:.
1134:.
1108:.
1082:.
1070:^
1054:24
1052:.
1048:.
1031:26
1029:.
1003:.
984:.
911:.
877:^
860:.
823:.
782:^
764:.
746:.
734:^
724:.
712:^
684:^
667:.
642:.
617:.
536:,
488:,
476:,
457:,
325:IP
55:.
1826:.
1798:.
1770:.
1744:.
1716:.
1688:.
1648:.
1629:.
1603:.
1577:.
1549:.
1521:.
1495:.
1462:.
1436:.
1430::
1421:,
1394:,
1388::
1359:.
1332::
1326:1
1264:.
1242:.
1228::
1205:.
1180:.
1168::
1145:.
1119:.
1093:.
1064:.
1014:.
988:.
935:.
922:.
871:.
834:.
776:.
750:.
728:.
678:.
653:.
628:.
528:.
120:,
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.