Nov 1 (Reuters) - CDW (CDW.O) reported third-quarter revenue below estimates on Wednesday on lower demand for the company's information ... their technology spend," the company said. CDW, which has products ranging from notebooks to data center services ...

CDW misses third-quarter revenue estimates as customers lower tech spending

By Robb M. Stewart Thomson Reuters' shares were up strongly Wednesday after the Canadian news and information provider announced a fresh $1 billion

Thomson Reuters Shares Rise With Earnings Beat, Buyback Plans

TORONTO (AP) — TORONTO (AP) — Thomson Reuters Corp. (TRI) on Wednesday reported third-quarter net income of $367 million. The Toronto-based company said it had profit of 80 cents per share. Earnings, adjusted for one-time gains and costs, were 82 cents per share.

Thomson Reuters: Q3 Earnings Snapshot

Thomson Reuters Corp. announced plans to invest more than US$100 million a year in generative artificial intelligence. The company said Wednesday it is building AI-powered capabilities for its products including a tool for its Westlaw Precision service that will launch Nov.

Thomson Reuters planning to invest more than US$100 million a year in generative AI

Thomson Reuters (TRI) delivered earnings and revenue surprises of 20.59% and 0.50%, respectively, for the quarter ended September 2023. Do the numbers hold clues to what lies ahead for the stock?

Thomson Reuters (TRI) Q3 Earnings Top Estimates

Earnings or losses for interim periods under IFRS reflect income taxes based on the estimated effective tax rates of each of the jurisdictions in which Thomson Reuters operates ... the core of the company's business information service product offerings.

Thomson Reuters Reports Third-Quarter 2023 Results

Thomson Reuters Corp on Wednesday reported higher-than-expected operating profit in the third quarter, helped by strong performance at its "Big 3" segments, Legal Professionals, Corporates and Tax & Accounting Professionals.

Thomson Reuters third-quarter profit beats estimates

The integration of the Westlaw Japan supports Thomson Reuters strategy to offer Japanese legal professionals a best-in-class content technology platform.

Thomson Reuters Completes Acquisition of Westlaw Japan Business

In our latest virtual conversation, HSToday Executive Editor Kristina Tanasichuk spoke with Heather Fischer, the Senior Advisor for Human Rights Crimes at Thomson Reuters Special Services and Minal Patel Davis,

Virtual Interview: New Partnership Creates Resources to Combat Human Trafficking

Microsoft Azure Customers Worldwide Now Gain Access to AU10TIX to Take Advantage of the Scalability, Reliability and Agility of Azure to Drive

AU10TIX ID Document & Biometric Verification Platform and Reusable ID Solution Now Available in the Microsoft Azure Marketplace

THOMSON REUTERS CORP

thomsonreuterscorp

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 7 of 15. F. The Times Files Suit It was only after this rapid adoption, along with reports of the value unlocked by these new technologies, that the Times claimed that OpenAI had “infringed its copyright[s]” and reached out to demand “commercial terms.” Compl. ¶¶ 54, 126. After months of discussions, the Times filed suit two days after Christmas, demanding “billions of dollars.” Id. ¶ 9. In its suit, the Times now claims ChatGPT will be the end of “independent journalism.” Id. ¶¶ 2, 47. Its chief complaint is that OpenAI “use[s]” news articles to build a service the world has never before seen that, according to the Times, might “steal [its] audiences.” Id. ¶¶ 2, 8. And its core legal theory is that copyright—a law singularly devoted to the “Progress of Science and useful Arts,” U.S. CONST. art. 1, § 8, cl. 8—should protect it from that technological innovation. Continue Reading Here. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 7 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 7 of 15. HackerNoon’s Legal PDF Series here here F. The Times Files Suit It was only after this rapid adoption, along with reports of the value unlocked by these new technologies, that the Times claimed that OpenAI had “infringed its copyright[s]” and reached out to demand “commercial terms.” Compl. ¶¶ 54, 126. After months of discussions, the Times filed suit two days after Christmas, demanding “billions of dollars.” Id. ¶ 9. In its suit, the Times now claims ChatGPT will be the end of “independent journalism.” Id. ¶¶ 2, 47. Its chief complaint is that OpenAI “use[s]” news articles to build a service the world has never before seen that, according to the Times, might “steal [its] audiences.” Id. ¶¶ 2, 8. And its core legal theory is that copyright—a law singularly devoted to the “Progress of Science and useful Arts,” U.S. CONST. art. 1, § 8, cl. 8—should protect it from that technological innovation. Continue Reading Here. Continue Reading Here . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

The New York Times Sued OpenAI Only After ChatGPT Became Popular

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 14 of 15. D. The “Misappropriation” Claim Is Preempted by the Copyright Act Count VI, for “misappropriation” under New York law, appears to raise two distinct theories. The first suggests that OpenAI engages in “unfair competition” by using “Times content to train models that produce informative text of the same general type and kind that [t]he Times produces.” Compl. ¶ 195 (the “Text Claim”). The second suggests that OpenAI harms the Times because ChatGPT can respond to user queries about Wirecutter recommendations. Id. ¶ 194 (the “Recommendations Claim”). Both are preempted by the Copyright Act. 17 U.S.C. § 301(a). 1. The Text Claim Fails The Text Claim is preempted by Section 301 of the Copyright Act, which “oust[s] the states from imposing any control of the area” governed by federal copyright law. In re Jackson, 972 F.3d 25, 42 (2d Cir. 2020). Preemption applies if two conditions are met: (1) the claim relates to “works of authorship . . . within [copyright’s] subject matter” (“subject matter” condition); and (2) the rights asserted are “equivalent to any of the exclusive rights within [copyright’s] general scope” (“general scope” condition). 17 U.S.C. § 301(a). Here, the subject matter condition is satisfied because the claim is based on OpenAI’s use of (1) Times articles, which are “literary works,” 17 U.S.C. § 102(a), and (2) facts from those articles which, while unprotectable, fall “within the subject matter of copyright for the purposes of [] preemption,” Barclays Capital Inc. v. Theflyonthewall.com, Inc., 650 F.3d 876, 893 (2d Cir. 2011). And the general scope condition is satisfied because the rights asserted are not “qualitatively different from a copyright [] claim.” In re Jackson, 972 F.3d at 43–44. One aspect of the Text Claim is based on the use of content to “train models,” Compl. ¶ 195, which is one of the stated bases of the copyright claim, id. ¶¶ 64, 161–62. The other is based on outputs that are “the same or similar to content published by [t]he Times,” id. ¶ 194–95, which is the other stated basis of the copyright claim, see id. ¶ 163. Because the Text Claim is based on the same allegations as the copyright claims, the claim is preempted.[47] 2. The Recommendations Claim Fails The Recommendations Claim focuses on Wirecutter, a Times-owned website that publishes articles to guide readers through difficult purchase decisions. See, e.g., Compl. ¶¶ 130, 134. Wirecutter makes “the vast majority of its revenue” through “commission[s],” which it earns when users click “affiliate links” in its articles and purchase a product. Id. ¶¶ 128–29. According to the Times, by complying with requests to identify Wirecutter-recommended products, ChatGPT reduces the need for users to “visit the [] Wirecutter article and click on [these] links,” id., thus “depriving [t]he Times of the opportunity to receive referral revenue,” id. ¶ 194. This claim is almost identical to the “misappropriation” claim the Second Circuit rejected in Barclays. 650 F.3d 876 (2d Cir. 2011). That case concerned stock recommendations which banks provided to paying clients as a mechanism for generating commissions. Id. at 880–82. The defendant “compile[d]” those recommendations and provided them “to their own subscribers for a fee.” Id. at 882. The banks brought a “misappropriation” claim, stressing that they “spend hundreds of millions of dollars annually” generating the recommendations, and that the defendant “seriously threatens their ability to justify [that] expense.” Id. at 881, 885–86. The Second Circuit first held that the claim satisfied the conditions for Copyright Act preemption: (1) the “facts of the Recommendations,” while unprotectable, fell within the subject matter of copyright; and (2) the banks claimed the right to control the reproduction of those recommendations, which falls within the “general scope” of copyright. Id. at 902. The Court, however, explained that its precedents recognized a “narrow” preemption exception for “hot news” claims endorsed by the Supreme Court in the Int’l News Serv. v. Associated Press (INS), 248 U.S. 215 (1918) case. Barclays, 650 F.3d at 896–98.[48] Accordingly, the merits of the claim depended on whether it fell into the “narrow” category of “INS-type non-preempted claims.” Id. at 902–03. The Circuit held it did not. First, while “INS-type” claims involve defendants who “sell [purloined] news as though the defendant itself had gathered it,” the Barclays defendant was “selling the information with specific attribution to the issuing [bank].” Id. (emphasis added). For that reason, there was no “meaningful difference” between the defendant’s republication of the recommendations and a “member[] of the traditional news media” reporting on information “with proper attribution.” Id. at 903–04. Second, INS-type claims concern defendants who appropriate material (e.g., news) that the defendants “acquire through efforts akin to reporting.” Id. at 903. But the banks’ claim sought “only to protect their Recommendations, something they create using their expertise and experience”—which brought the claim closer to copyright’s exclusive domain. Id. For those reasons, the defendant’s “service—which collects, summarizes, and disseminates the news of the [banks’] Recommendations—[was] not the kind of ‘INS-like’ product that could support a non-preempted cause of action for misappropriation.” Id. at 905. The Recommendations Claim fails for the same reasons. First, it satisfies the preemption conditions: it (1) relates to facts within copyright’s “subject matter;” and (2) asserts the right to control reproduction of those facts, which falls within copyright’s “general scope.” Id. at 902. Second, to the extent the Times has attempted to plead an “INS-type non-preempted claim[]” for misappropriation, id. at 902, it has failed to do so for the reasons discussed in Barclays. OpenAI is not “selling the Recommendation[s] ‘as its own.’” Id. at 903. It provides “specific attribution to” Wirecutter, which means this is not an INS claim. Id. Moreover, Wirecutter recommendations are not facts that the Times “acquire[s] through efforts akin to reporting;” rather, the Times “seek[s] only to protect [its] Recommendations, [which it] create[s] using [its] expertise and experience.” Id. The claim is therefore governed exclusively by copyright law, under which the quantum of the Times’s “investment” is irrelevant to its right to monopolize the results. Compl. ¶ 196; Moody’s, 808 F.2d at 207 (“[T]o grant copyright protection based merely on the ‘sweat of the author’s brow’ would risk putting large areas of factual research material off limits and threaten the public’s unrestrained access to information.”); Barclays, 650 F.3d at 886 (rejecting claim despite allegation that defendant “seriously threatens [banks’] ability to justify the expense of maintaining [] extensive research operations”). Continue Reading Here. [47] Financial Information, Inc. v. Moody’s Investors Serv., Inc., 808 F.2d 204, 206, 208–09 (2d Cir. 1986) (misappropriation claim based on allegation that defendant “copied 40–50% of [plaintiff’s] information” preempted). [48] INS concerned a news service that “lift[ed] factual stories from AP bulletins and sen[t] them by wire to INS papers” for republication without attribution to the AP. Barclays, 650 F.3d at 894, 896–98 (quoting INS). About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 14 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 14 of 15. HackerNoon’s Legal PDF Series here here D. The “Misappropriation” Claim Is Preempted by the Copyright Act Count VI, for “misappropriation” under New York law, appears to raise two distinct theories. The first suggests that OpenAI engages in “unfair competition” by using “Times content to train models that produce informative text of the same general type and kind that [t]he Times produces.” Compl. ¶ 195 (the “Text Claim”). The second suggests that OpenAI harms the Times because ChatGPT can respond to user queries about Wirecutter recommendations. Id. ¶ 194 (the “Recommendations Claim”). Both are preempted by the Copyright Act. 17 U.S.C. § 301(a). 1. The Text Claim Fails 1. The Text Claim Fails The Text Claim is preempted by Section 301 of the Copyright Act, which “oust[s] the states from imposing any control of the area” governed by federal copyright law. In re Jackson, 972 F.3d 25, 42 (2d Cir. 2020). Preemption applies if two conditions are met: (1) the claim relates to “works of authorship . . . within [copyright’s] subject matter” (“subject matter” condition); and (2) the rights asserted are “equivalent to any of the exclusive rights within [copyright’s] general scope” (“general scope” condition). 17 U.S.C. § 301(a). Here, the subject matter condition is satisfied because the claim is based on OpenAI’s use of (1) Times articles, which are “literary works,” 17 U.S.C. § 102(a), and (2) facts from those articles which, while unprotectable, fall “within the subject matter of copyright for the purposes of [] preemption,” Barclays Capital Inc. v. Theflyonthewall.com, Inc., 650 F.3d 876, 893 (2d Cir. 2011). And the general scope condition is satisfied because the rights asserted are not “qualitatively different from a copyright [] claim.” In re Jackson, 972 F.3d at 43–44. One aspect of the Text Claim is based on the use of content to “train models,” Compl. ¶ 195, which is one of the stated bases of the copyright claim, id. ¶¶ 64, 161–62. The other is based on outputs that are “the same or similar to content published by [t]he Times,” id. ¶ 194–95, which is the other stated basis of the copyright claim, see id. ¶ 163. Because the Text Claim is based on the same allegations as the copyright claims, the claim is preempted.[47] 2. The Recommendations Claim Fails 2. The Recommendations Claim Fails The Recommendations Claim focuses on Wirecutter, a Times-owned website that publishes articles to guide readers through difficult purchase decisions. See, e.g., Compl. ¶¶ 130, 134. Wirecutter makes “the vast majority of its revenue” through “commission[s],” which it earns when users click “affiliate links” in its articles and purchase a product. Id. ¶¶ 128–29. According to the Times, by complying with requests to identify Wirecutter-recommended products, ChatGPT reduces the need for users to “visit the [] Wirecutter article and click on [these] links,” id., thus “depriving [t]he Times of the opportunity to receive referral revenue,” id. ¶ 194. This claim is almost identical to the “misappropriation” claim the Second Circuit rejected in Barclays. 650 F.3d 876 (2d Cir. 2011). That case concerned stock recommendations which banks provided to paying clients as a mechanism for generating commissions. Id. at 880–82. The defendant “compile[d]” those recommendations and provided them “to their own subscribers for a fee.” Id. at 882. The banks brought a “misappropriation” claim, stressing that they “spend hundreds of millions of dollars annually” generating the recommendations, and that the defendant “seriously threatens their ability to justify [that] expense.” Id. at 881, 885–86. The Second Circuit first held that the claim satisfied the conditions for Copyright Act preemption: (1) the “facts of the Recommendations,” while unprotectable, fell within the subject matter of copyright; and (2) the banks claimed the right to control the reproduction of those recommendations, which falls within the “general scope” of copyright. Id. at 902. The Court, however, explained that its precedents recognized a “narrow” preemption exception for “hot news” claims endorsed by the Supreme Court in the Int’l News Serv. v. Associated Press (INS), 248 U.S. 215 (1918) case. Barclays, 650 F.3d at 896–98.[48] Accordingly, the merits of the claim depended on whether it fell into the “narrow” category of “INS-type non-preempted claims.” Id. at 902–03. The Circuit held it did not. First, while “INS-type” claims involve defendants who “sell [purloined] news as though the defendant itself had gathered it,” the Barclays defendant was “selling the information with specific attribution to the issuing [bank].” Id. (emphasis added). For that reason, there was no “meaningful difference” between the defendant’s republication of the recommendations and a “member[] of the traditional news media” reporting on information “with proper attribution.” Id. at 903–04. Second, INS-type claims concern defendants who appropriate material (e.g., news) that the defendants “acquire through efforts akin to reporting.” Id. at 903. But the banks’ claim sought “only to protect their Recommendations, something they create using their expertise and experience”—which brought the claim closer to copyright’s exclusive domain. Id. For those reasons, the defendant’s “service—which collects, summarizes, and disseminates the news of the [banks’] Recommendations—[was] not the kind of ‘INS-like’ product that could support a non-preempted cause of action for misappropriation.” Id. at 905. The Recommendations Claim fails for the same reasons. First, it satisfies the preemption conditions: it (1) relates to facts within copyright’s “subject matter;” and (2) asserts the right to control reproduction of those facts, which falls within copyright’s “general scope.” Id. at 902. Second, to the extent the Times has attempted to plead an “INS-type non-preempted claim[]” for misappropriation, id. at 902, it has failed to do so for the reasons discussed in Barclays. OpenAI is not “selling the Recommendation[s] ‘as its own.’” Id. at 903. It provides “specific attribution to” Wirecutter, which means this is not an INS claim. Id. Moreover, Wirecutter recommendations are not facts that the Times “acquire[s] through efforts akin to reporting;” rather, the Times “seek[s] only to protect [its] Recommendations, [which it] create[s] using [its] expertise and experience.” Id. The claim is therefore governed exclusively by copyright law, under which the quantum of the Times’s “investment” is irrelevant to its right to monopolize the results. Compl. ¶ 196; Moody’s, 808 F.2d at 207 (“[T]o grant copyright protection based merely on the ‘sweat of the author’s brow’ would risk putting large areas of factual research material off limits and threaten the public’s unrestrained access to information.”); Barclays, 650 F.3d at 886 (rejecting claim despite allegation that defendant “seriously threatens [banks’] ability to justify the expense of maintaining [] extensive research operations”). Continue Reading Here. Continue Reading Here . Here [47] Financial Information, Inc. v. Moody’s Investors Serv., Inc., 808 F.2d 204, 206, 208–09 (2d Cir. 1986) (misappropriation claim based on allegation that defendant “copied 40–50% of [plaintiff’s] information” preempted). [48] INS concerned a news service that “lift[ed] factual stories from AP bulletins and sen[t] them by wire to INS papers” for republication without attribution to the AP. Barclays, 650 F.3d at 894, 896–98 (quoting INS). About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

OpenAI Harms the NYT Because of ChatGPT's Response to User Queries About Wirecutter Recommendations

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 9 of 15. III. LEGAL STANDARD “[A] complaint must contain sufficient factual matter, accepted as true, to state a claim to relief that is plausible on its face.” Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). “[C]onclusory allegations or legal conclusions masquerading as fact[s]” do not suffice. Arcesium, LLC v. Advent Software, Inc., No. 20-cv-04389, 2021 WL 1225446, at *5 (S.D.N.Y. Mar. 31, 2021). Continue Reading Here. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 9 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 9 of 15. HackerNoon’s Legal PDF Series here here III. LEGAL STANDARD “[A] complaint must contain sufficient factual matter, accepted as true, to state a claim to relief that is plausible on its face.” Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). “[C]onclusory allegations or legal conclusions masquerading as fact[s]” do not suffice. Arcesium, LLC v. Advent Software, Inc., No. 20-cv-04389, 2021 WL 1225446, at *5 (S.D.N.Y. Mar. 31, 2021). Continue Reading Here. Continue Reading Here . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

OpenAI Presents Legal Standard in NYT Dismissal Argument

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 2 of 15. II. BACKGROUND A. OpenAI’s Pioneering Research OpenAI was founded in 2015 to “advance digital intelligence in the way that is most likely to benefit humanity as a whole.” Compl. ¶ 56. It entered the field of “natural language processing” (NLP), which includes the development of statistical tools called “language models.”[15] These models can “predict[] words that are likely to follow a given string of text” based on statistics derived from a body of text—much like a weather model can predict the rain using statistics derived from historical weather data. Compl. ¶ 75. By 2015, research had already unlocked “substantial progress” on “tasks such as reading comprehension” and “question answering.”[16] Those early models, however, were “brittle” and “narrow.”[17] Researchers built them by “manually creat[ing] and label[ling]” datasets to “demonstrate[e] correct behavior”—like sets of English-to-French text translations—and using that data to “train a system to imitate [that] behavior[].” GPT-2 Paper at 1, 3. The resulting models, while impressive, could only carry out the specific tasks demonstrated by the training data. Id.; GPT-3 Paper at 3 (“need for task-specific datasets” was “a major limitation”). “To be broadly useful” to ordinary people, language models needed the ability to “seamlessly mix together or switch between many tasks and skills” without being specifically trained to carry out each task. GPT-3 Paper at 4. In other words, the models needed to be “competent generalists,” not “narrow experts.” GPT-2 Paper at 1. OpenAI’s researchers set out to solve that complex, scientific problem. In 2019, they posited that the way to build more capable, generalist models was to use “as large and diverse a dataset as possible [] to collect natural language demonstrations of tasks in as varied of domains and contexts as possible.” GPT-2 Paper at 3. The hypothesis was that “[]training at a large enough scale [might] offer a ‘natural’ broad distribution of tasks implicitly contained in predicting the text itself.” GPT-3 Paper at 40. So instead of training its models “on a single domain of text,” OpenAI chose to use a richer and more diverse source: the Internet. GPT-2 Paper at 3. OpenAI’s researchers identified text from webpages whose URLs had been publicly shared on a social media platform. Id. This became a dataset called “WebText,” which OpenAI used to train a model called “GPT-2.” Id.; see also Compl. ¶ 85. WebText contained a wide array of text from internet forums, restaurant reviews, recipe websites, blogs, shopping websites, dictionaries, medical websites, how-to pages, and more.[18] The dataset was so diverse that even though Times content represented only a tiny fraction of the data, the “NYTimes” was one of the “top 15 domains by volume” in the collection. See GPT-2 Model Card. This happened not because OpenAI believed Times articles are more “valu[able]” than other content, contra Compl. ¶ 2 (suggesting OpenAI intentionally “gave Times content particular emphasis”), but because of the frequency with which certain social media users shared links to the Times’s content, see GPT-2 Paper at 3. The results of this sophisticated research were impressive. The GPT-2 model proved able to answer trivia questions and perform higher-function tasks like “resolv[ing] ambiguities in text.” GPT-2 Paper at 6–7. The model even showed a “surprising” ability to translate French to English, even though OpenAI had “deliberately removed non-English webpages” from the training dataset. Id. at 7. These research results were “exciting” not only because of the model’s capability, but because they scientifically confirmed that the ability to “perform commonsense reasoning” increased dramatically with the size and diversity of the training data. Id. at 6 (Figure 3). Continue Reading Here. [15] Sébastien Bubeck, et al., Sparks of Artificial General Intelligence: Early Experiments with GPT-4 at 4, 98 (Apr. 13, 2023), https://arxiv.org/pdf/2303.12712.pdf (“Bubeck Paper”); Compl. ¶¶ 71, 91 nn.9 & 24 (citing articles). By “refer[ing] [to these documents] in [its] complaint,” the Times incorporated them by reference. DiFolco v. MSNBC Cable L.L.C., 622 F.3d 104, 111–12 (2d Cir. 2010). [16] OpenAI, Language Models are Few-Shot Learners at 3 (July 22, 2020), https://arxiv.org/pdf/2005.14165.pdf (“GPT-3 Paper”); see also Compl. ¶¶ 86, 90 & nn.18, 22 (citing and quoting this paper). [17] OpenAI, Language Models are Unsupervised Multitask Learners at 1 (Feb. 14, 2019), https://cdn.openai.com/better-language-models/language_models_are_unsupervised_multitask_learners.pdf (“GPT-2 Paper”); see also Compl. ¶ 85 n.15 (citing and quoting this paper). [18] See OpenAI, GPT-2 Model Card, Github, https://github.com/openai/gpt-2/blob/master/model_card.md (last updated Nov. 2019) (“GPT-2 Model Card”); see also Compl. ¶ 85 nn. 14, 16, 17 (citing and quoting this source). About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 2 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 2 of 15. HackerNoon’s Legal PDF Series here here II. BACKGROUND A. OpenAI’s Pioneering Research OpenAI was founded in 2015 to “advance digital intelligence in the way that is most likely to benefit humanity as a whole.” Compl. ¶ 56. It entered the field of “natural language processing” (NLP), which includes the development of statistical tools called “language models.”[15] These models can “predict[] words that are likely to follow a given string of text” based on statistics derived from a body of text—much like a weather model can predict the rain using statistics derived from historical weather data. Compl. ¶ 75. By 2015, research had already unlocked “substantial progress” on “tasks such as reading comprehension” and “question answering.”[16] Those early models, however, were “brittle” and “narrow.”[17] Researchers built them by “manually creat[ing] and label[ling]” datasets to “demonstrate[e] correct behavior”—like sets of English-to-French text translations—and using that data to “train a system to imitate [that] behavior[].” GPT-2 Paper at 1, 3. The resulting models, while impressive, could only carry out the specific tasks demonstrated by the training data. Id.; GPT-3 Paper at 3 (“need for task-specific datasets” was “a major limitation”). “To be broadly useful” to ordinary people, language models needed the ability to “seamlessly mix together or switch between many tasks and skills” without being specifically trained to carry out each task. GPT-3 Paper at 4. In other words, the models needed to be “competent generalists,” not “narrow experts.” GPT-2 Paper at 1. OpenAI’s researchers set out to solve that complex, scientific problem. In 2019, they posited that the way to build more capable, generalist models was to use “as large and diverse a dataset as possible [] to collect natural language demonstrations of tasks in as varied of domains and contexts as possible.” GPT-2 Paper at 3. The hypothesis was that “[]training at a large enough scale [might] offer a ‘natural’ broad distribution of tasks implicitly contained in predicting the text itself.” GPT-3 Paper at 40. So instead of training its models “on a single domain of text,” OpenAI chose to use a richer and more diverse source: the Internet. GPT-2 Paper at 3. OpenAI’s researchers identified text from webpages whose URLs had been publicly shared on a social media platform. Id. This became a dataset called “WebText,” which OpenAI used to train a model called “GPT-2.” Id.; see also Compl. ¶ 85. WebText contained a wide array of text from internet forums, restaurant reviews, recipe websites, blogs, shopping websites, dictionaries, medical websites, how-to pages, and more.[18] The dataset was so diverse that even though Times content represented only a tiny fraction of the data, the “NYTimes” was one of the “top 15 domains by volume” in the collection. See GPT-2 Model Card. This happened not because OpenAI believed Times articles are more “valu[able]” than other content, contra Compl. ¶ 2 (suggesting OpenAI intentionally “gave Times content particular emphasis”), but because of the frequency with which certain social media users shared links to the Times’s content, see GPT-2 Paper at 3. The results of this sophisticated research were impressive. The GPT-2 model proved able to answer trivia questions and perform higher-function tasks like “resolv[ing] ambiguities in text.” GPT-2 Paper at 6–7. The model even showed a “surprising” ability to translate French to English, even though OpenAI had “deliberately removed non-English webpages” from the training dataset. Id. at 7. These research results were “exciting” not only because of the model’s capability, but because they scientifically confirmed that the ability to “perform commonsense reasoning” increased dramatically with the size and diversity of the training data. Id. at 6 (Figure 3). Continue Reading Here. Continue Reading Here . Here [15] Sébastien Bubeck, et al., Sparks of Artificial General Intelligence: Early Experiments with GPT-4 at 4, 98 (Apr. 13, 2023), https://arxiv.org/pdf/2303.12712.pdf (“Bubeck Paper”); Compl. ¶¶ 71, 91 nn.9 & 24 (citing articles). By “refer[ing] [to these documents] in [its] complaint,” the Times incorporated them by reference. DiFolco v. MSNBC Cable L.L.C., 622 F.3d 104, 111–12 (2d Cir. 2010). [16] OpenAI, Language Models are Few-Shot Learners at 3 (July 22, 2020), https://arxiv.org/pdf/2005.14165.pdf (“GPT-3 Paper”); see also Compl. ¶¶ 86, 90 & nn.18, 22 (citing and quoting this paper). [17] OpenAI, Language Models are Unsupervised Multitask Learners at 1 (Feb. 14, 2019), https://cdn.openai.com/better-language-models/language_models_are_unsupervised_multitask_learners.pdf (“GPT-2 Paper”); see also Compl. ¶ 85 n.15 (citing and quoting this paper). [18] See OpenAI, GPT-2 Model Card, Github, https://github.com/openai/gpt-2/blob/master/model_card.md (last updated Nov. 2019) (“GPT-2 Model Card”); see also Compl. ¶ 85 nn. 14, 16, 17 (citing and quoting this source). About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

ChatGPT's Predecessor Could Translate French Even Though It Was Not Trained to Do So

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 3 of 15. B. The Key to Generalist Language Models: Scale So OpenAI’s research took the logical next step: “scaling up [] model size” by increasing both the “size” and “diversity” of the training data. See GPT-3 Paper at 6. To build its next generation of models, OpenAI’s researchers gathered a more robust set of data in part by “expand[ing]” WebText database into a new version called “WebText2,” which included material shared over a longer period of time. Id. at 8–9. They also used a filtered version of Common Crawl, a repository of data collected by a non-profit research organization representing “a copy of the Internet.” Compl. ¶ 88. OpenAI disclosed all of this no later than July 22, 2020. GPT-3 Paper at 8. At the time, it was common knowledge that WebText2 and Common Crawl included numerous articles published by the Times.[19] The result of this simple act of “scaling up” the training data was, as the Times reported at the time, “mind blowing.”[20] The new GPT-3 model was “by far the most powerful ‘language model’ ever created.” Manjoo, supra note 20. It could conduct “on-the-fly reasoning” and “unscrambl[e] words, perform[] arithmetic, and us[e] novel words in a sentence after seeing them defined only once.” GPT-3 Paper at 5. Increasing the scale of training led to a surprising jump in ability. See, e.g., id. at 22 (Figure 3.10). According to the Times’s reporting, GPT-3 showed that “[m]achines are gaining the ability to write.” Manjoo, supra note 20. Within days, developers began to use it to build unprecedented tools. Id. (“service that responds to email on your behalf”). The “key advance,” as the Times reported, was “GPT-3’s flexibility.” Id. And the reason it was flexible was OpenAI’s decision to “scal[e] up” the “size” of the training data. GPT-3 Paper at 6. The amount of data needed was staggering. Compl. ¶ 85. But it was that “unprecedented scale” that allowed the model to internalize not only a “map of human language,” but achieve a level of adaptability—and “emergent” intelligence—that “no one thought possible.”[21] Continue Reading Here. [19] Compl. ¶ 88 (citing a “2019 snapshot of Common Crawl”); GPT-2 Model Card (noting prevalence of Times content in WebText); GPT-3 Paper at 8 (noting WebText2 is an “expanded version of the WebText dataset”). [20] Farhad Manjoo, How Do You Know a Human Wrote This?, N.Y. Times (July 29, 2020), https://www.nytimes.com/2020/07/29/opinion/gpt-3-ai-automation.html. The Court may take judicial notice of this and other news articles cited in this Motion. In re UBS Auction Rate Sec. Litig., No. 08-cv-2967, 2010 WL 2541166, at *10-12 (S.D.N.Y. June 10, 2010) (judicial notice of “several news items” including Wall Street Journal articles). [21] Cade Metz, Meet GPT-3. It Has Learned to Code (and Blog and Argue), N.Y. Times (Nov. 24, 2020), https://www.nytimes.com/2020/11/24/science/artificial-intelligence-ai-gpt3.html. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 3 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 3 of 15. HackerNoon’s Legal PDF Series here here B. The Key to Generalist Language Models: Scale So OpenAI’s research took the logical next step: “scaling up [] model size” by increasing both the “size” and “diversity” of the training data. See GPT-3 Paper at 6. To build its next generation of models, OpenAI’s researchers gathered a more robust set of data in part by “expand[ing]” WebText database into a new version called “WebText2,” which included material shared over a longer period of time. Id. at 8–9. They also used a filtered version of Common Crawl, a repository of data collected by a non-profit research organization representing “a copy of the Internet.” Compl. ¶ 88. OpenAI disclosed all of this no later than July 22, 2020. GPT-3 Paper at 8. At the time, it was common knowledge that WebText2 and Common Crawl included numerous articles published by the Times.[19] The result of this simple act of “scaling up” the training data was, as the Times reported at the time, “mind blowing.”[20] The new GPT-3 model was “by far the most powerful ‘language model’ ever created.” Manjoo, supra note 20. It could conduct “on-the-fly reasoning” and “unscrambl[e] words, perform[] arithmetic, and us[e] novel words in a sentence after seeing them defined only once.” GPT-3 Paper at 5. Increasing the scale of training led to a surprising jump in ability. See, e.g., id. at 22 (Figure 3.10). According to the Times’s reporting, GPT-3 showed that “[m]achines are gaining the ability to write.” Manjoo, supra note 20. Within days, developers began to use it to build unprecedented tools. Id. (“service that responds to email on your behalf”). The “key advance,” as the Times reported, was “GPT-3’s flexibility.” Id. And the reason it was flexible was OpenAI’s decision to “scal[e] up” the “size” of the training data. GPT-3 Paper at 6. The amount of data needed was staggering. Compl. ¶ 85. But it was that “unprecedented scale” that allowed the model to internalize not only a “map of human language,” but achieve a level of adaptability—and “emergent” intelligence—that “no one thought possible.”[21] Continue Reading Here. Continue Reading Here . Here [19] Compl. ¶ 88 (citing a “2019 snapshot of Common Crawl”); GPT-2 Model Card (noting prevalence of Times content in WebText); GPT-3 Paper at 8 (noting WebText2 is an “expanded version of the WebText dataset”). [20] Farhad Manjoo, How Do You Know a Human Wrote This?, N.Y. Times (July 29, 2020), https://www.nytimes.com/2020/07/29/opinion/gpt-3-ai-automation.html. The Court may take judicial notice of this and other news articles cited in this Motion. In re UBS Auction Rate Sec. Litig., No. 08-cv-2967, 2010 WL 2541166, at *10-12 (S.D.N.Y. June 10, 2010) (judicial notice of “several news items” including Wall Street Journal articles). [21] Cade Metz, Meet GPT-3. It Has Learned to Code (and Blog and Argue), N.Y. Times (Nov. 24, 2020), https://www.nytimes.com/2020/11/24/science/artificial-intelligence-ai-gpt3.html. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

ChatGPT's Predecessor Was Trained on 'Filtered' Internet Data from an Open Source Repository

Because It Cannot Sue ChatGPT For Mimicry, the NYT Lawsuit Focuses on AI Hallucination Instead

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 6 of 15. E. OpenAI’s API and ChatGPT Today, there are two primary ways to interact with OpenAI’s models. The first is an application programming interface (“API”) that allows developers to prompt the models directly with computer code. Compl. ¶ 62. Developers must specify a number of parameters, including their desired “Model” version and “Temperature” (i.e., the degree of randomness). See, e.g., id. ¶ 140. Developers can supply a “System” prompt, which tells the model how to behave, e.g., “You are a thespian who speaks only in iambic pentameter.” Id. OpenAI also provides a “Playground” for developers to experiment with the models. See id. The second is ChatGPT, a consumer-friendly platform—accessible for free at chat.openai.com—to “chat” with an OpenAI model through a user interface.[24] Its “Browse with Bing” feature enables ChatGPT to fetch recent information about events that occurred after the models’ training “cutoff,” using the Bing search engine. Compl. ¶¶ 72, 108, 112. Using it is the equivalent of instructing a research assistant to use the Internet to look up a fact she does not yet know. The feature goes hand-in-hand with others that collectively amplify ChatGPT’s utility. See Bubeck Paper at 45 (describing ability to “coordinate a dinner” using “APIs to retrieve information about the user’s calendar, coordinate with other people over email, [and] book the dinner”). After its release in November 2022, ChatGPT became an “instant viral sensation.” Compl. ¶ 61. According to the Complaint, by April 2023, the service had 173 million users globally who rely on ChatGPT to “craft clearer communications, accelerate coding tasks, rapidly explore answers to complex business questions, assist with creative work, and much more.”[25] Continue Reading Here. [24] Compl. ¶¶ 62, 102; OpenAI Blog, Introducing ChatGPT Plus (Feb. 1, 2023), https://openai.com/blog/chatgpt-plus; see also Compl. ¶ 149 n.41 (citing this page). [25] Compl. ¶ 149; OpenAI Blog, Introducing ChatGPT Enterprise (Aug. 28, 2023), https://openai.com/blog/introducing-chatgpt-enterprise; see also Compl. ¶¶ 63, 149 & nn.5 & 42 (citing this article); About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 6 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 6 of 15. HackerNoon’s Legal PDF Series here here E. OpenAI’s API and ChatGPT Today, there are two primary ways to interact with OpenAI’s models. The first is an application programming interface (“API”) that allows developers to prompt the models directly with computer code. Compl. ¶ 62. Developers must specify a number of parameters, including their desired “Model” version and “Temperature” (i.e., the degree of randomness). See, e.g., id. ¶ 140. Developers can supply a “System” prompt, which tells the model how to behave, e.g., “You are a thespian who speaks only in iambic pentameter.” Id. OpenAI also provides a “Playground” for developers to experiment with the models. See id. The second is ChatGPT, a consumer-friendly platform—accessible for free at chat.openai.com—to “chat” with an OpenAI model through a user interface.[24] Its “Browse with Bing” feature enables ChatGPT to fetch recent information about events that occurred after the models’ training “cutoff,” using the Bing search engine. Compl. ¶¶ 72, 108, 112. Using it is the equivalent of instructing a research assistant to use the Internet to look up a fact she does not yet know. The feature goes hand-in-hand with others that collectively amplify ChatGPT’s utility. See Bubeck Paper at 45 (describing ability to “coordinate a dinner” using “APIs to retrieve information about the user’s calendar, coordinate with other people over email, [and] book the dinner”). After its release in November 2022, ChatGPT became an “instant viral sensation.” Compl. ¶ 61. According to the Complaint, by April 2023, the service had 173 million users globally who rely on ChatGPT to “craft clearer communications, accelerate coding tasks, rapidly explore answers to complex business questions, assist with creative work, and much more.”[25] Continue Reading Here. Continue Reading Here . Here [24] Compl. ¶¶ 62, 102; OpenAI Blog, Introducing ChatGPT Plus (Feb. 1, 2023), https://openai.com/blog/chatgpt-plus; see also Compl. ¶ 149 n.41 (citing this page). [25] Compl. ¶ 149; OpenAI Blog, Introducing ChatGPT Enterprise (Aug. 28, 2023), https://openai.com/blog/introducing-chatgpt-enterprise; see also Compl. ¶¶ 63, 149 & nn.5 & 42 (citing this article); About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

OpenAI Provides a “Playground” for Developers to Experiment With

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 4 of 15. C. Reliance on Longstanding Fair Use Principles By July 2020, OpenAI had disclosed that Times articles were a tiny part of the diverse datasets that had been used to train these language models. And according to the Complaint, by the time GPT-3 was released in mid-2020, OpenAI had already established itself as a “commercial enterprise.” Compl. ¶ 57. The Times itself reported in 2020 that “OpenAI plans to sell access to GPT-3 via the internet, turning it into a widely used commercial product.” Metz, supra note 21. While its reporters joked they “might be put out to pasture by a machine,” Manjoo, supra note 20, the Times never accused OpenAI of violating copyright law. Instead, the Times enthusiastically and factually reported that the technology could be “enormously useful” and “open[] the door to a wide range of new possibilities.” Manjoo, supra note 20; Metz, supra note 21. Indeed, it has long been clear that the non-consumptive use of copyrighted material (like large language model training) is protected by fair use—a doctrine as important to the Times itself as it is to the American technology industry.[22] Since Congress codified that doctrine in 1976, see H.R. Rep. No. 94-1476, at 65–66 (1976) (courts should “adapt” defense to “rapid technological change”), courts have used it to protect useful innovations like home video recording, internet search, book search tools, reuse of software APIs, and many others.[23] These precedents reflect the foundational principle that copyright law exists to control the dissemination of works in the marketplace—not to grant authors “absolute control” over all uses of their works. Google Books, 804 F.3d at 212. Copyright is not a veto right over transformative technologies that leverage existing works internally—i.e., without disseminating them—to new and useful ends, thereby furthering copyright’s basic purpose without undercutting authors’ ability to sell their works in the marketplace. See supra note 23. And it is the “basic purpose” of fair use to “keep [the] copyright monopoly within [these] lawful bounds.” Oracle, 141 S. Ct. at 1198. OpenAI and scores of other developers invested billions of dollars, and the efforts of some of the world’s most capable minds, based on these clear and longstanding principles. Continue Reading Here. [22] See, e.g., Edmund White, In ‘The Talented Mr. Ripley,’ A Shape-Shifting Protagonist Who’s Up to No Good, N.Y. Times Style Magazine (Mar. 24, 2021), https://www.nytimes.com/2021/03/24/t-magazine/talented-mr-ripleypatricia-highsmith.html (including continuous 200-word excerpt from published novel). [23] Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 454–55 (1984); Kelly v. Arriba Soft Corp., 336 F.3d 811, 818–22 (9th Cir. 2003); Google Books, 804 F.3d at 209; Oracle, 141 S. Ct. at 1209 (2021). About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 4 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 4 of 15. HackerNoon’s Legal PDF Series here here C. Reliance on Longstanding Fair Use Principles By July 2020, OpenAI had disclosed that Times articles were a tiny part of the diverse datasets that had been used to train these language models. And according to the Complaint, by the time GPT-3 was released in mid-2020, OpenAI had already established itself as a “commercial enterprise.” Compl. ¶ 57. The Times itself reported in 2020 that “OpenAI plans to sell access to GPT-3 via the internet, turning it into a widely used commercial product.” Metz, supra note 21. While its reporters joked they “might be put out to pasture by a machine,” Manjoo, supra note 20, the Times never accused OpenAI of violating copyright law. Instead, the Times enthusiastically and factually reported that the technology could be “enormously useful” and “open[] the door to a wide range of new possibilities.” Manjoo, supra note 20; Metz, supra note 21. Indeed, it has long been clear that the non-consumptive use of copyrighted material (like large language model training) is protected by fair use—a doctrine as important to the Times itself as it is to the American technology industry.[22] Since Congress codified that doctrine in 1976, see H.R. Rep. No. 94-1476, at 65–66 (1976) (courts should “adapt” defense to “rapid technological change”), courts have used it to protect useful innovations like home video recording, internet search, book search tools, reuse of software APIs, and many others.[23] These precedents reflect the foundational principle that copyright law exists to control the dissemination of works in the marketplace—not to grant authors “absolute control” over all uses of their works. Google Books, 804 F.3d at 212. Copyright is not a veto right over transformative technologies that leverage existing works internally—i.e., without disseminating them—to new and useful ends, thereby furthering copyright’s basic purpose without undercutting authors’ ability to sell their works in the marketplace. See supra note 23. And it is the “basic purpose” of fair use to “keep [the] copyright monopoly within [these] lawful bounds.” Oracle, 141 S. Ct. at 1198. OpenAI and scores of other developers invested billions of dollars, and the efforts of some of the world’s most capable minds, based on these clear and longstanding principles. Continue Reading Here. Continue Reading Here . Here [22] See, e.g., Edmund White, In ‘The Talented Mr. Ripley,’ A Shape-Shifting Protagonist Who’s Up to No Good, N.Y. Times Style Magazine (Mar. 24, 2021), https://www.nytimes.com/2021/03/24/t-magazine/talented-mr-ripleypatricia-highsmith.html (including continuous 200-word excerpt from published novel). [23] Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 454–55 (1984); Kelly v. Arriba Soft Corp., 336 F.3d 811, 818–22 (9th Cir. 2003); Google Books, 804 F.3d at 209; Oracle, 141 S. Ct. at 1209 (2021). About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

New York Times Articles Are a 'Tiny Part' of ChatGPT's Training Data

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 5 of 15. D. GPT-3.5 and GPT-4 OpenAI built on that success by researching, developing, and releasing two other models— GPT-3.5 (in 2022) and GPT-4 (in 2023)—which triggered the AI revolution that we are living through today. GPT-4 “can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more.” Bubeck Paper at 1. Researchers already use these models for a growing list of astonishingly productive ends. See supra 1. Continue Reading Here. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 5 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 5 of 15. HackerNoon’s Legal PDF Series here here D. GPT-3.5 and GPT-4 OpenAI built on that success by researching, developing, and releasing two other models— GPT-3.5 (in 2022) and GPT-4 (in 2023)—which triggered the AI revolution that we are living through today. GPT-4 “can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more.” Bubeck Paper at 1. Researchers already use these models for a growing list of astonishingly productive ends. See supra 1. Continue Reading Here. Continue Reading Here . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

OpenAI's GPT-3.5 and GPT-4 'Triggered' the AI Revolution 

The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 10 of 15. IV. ARGUMENT This Motion seeks dismissal of four claims. (1) OpenAI seeks partial dismissal of Count I (Direct Copyright Infringement) to the extent it is based on acts of reproduction that occurred more than three years before this action. Infra Section IV(A). (2) OpenAI seeks full dismissal of Count IV (Contributory Infringement) for failure to allege that it had actual knowledge of the specific acts of direct infringement alleged. Infra Section IV(B). (3) OpenAI seeks full dismissal of Count V (Copyright Management Information or “CMI” Removal) for several reasons, including failure to identify the CMI at issue, failure to allege OpenAI “remove[d]” CMI from any datasets or outputs, failure to allege “distribution,” and failure to plead facts that suggest that OpenAI acted with scienter. Infra Section IV(C). And (4) OpenAI seeks full dismissal of Count VI (Unfair Competition by Misappropriation) on grounds of Copyright Act preemption. Infra Section IV(D). Continue Reading Here. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings.
This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series. You can jump to any part in this filing here. This part is 10 of 15. The New York Times Company v. OpenAI Update Court Filing, retrieved on February 26, 2024 is part of HackerNoon’s Legal PDF Series . You can jump to any part in this filing here . This part is 10 of 15. HackerNoon’s Legal PDF Series here here IV. ARGUMENT This Motion seeks dismissal of four claims. (1) OpenAI seeks partial dismissal of Count I (Direct Copyright Infringement) to the extent it is based on acts of reproduction that occurred more than three years before this action. Infra Section IV(A). (2) OpenAI seeks full dismissal of Count IV (Contributory Infringement) for failure to allege that it had actual knowledge of the specific acts of direct infringement alleged. Infra Section IV(B). (3) OpenAI seeks full dismissal of Count V (Copyright Management Information or “CMI” Removal) for several reasons, including failure to identify the CMI at issue, failure to allege OpenAI “remove[d]” CMI from any datasets or outputs, failure to allege “distribution,” and failure to plead facts that suggest that OpenAI acted with scienter. Infra Section IV(C). And (4) OpenAI seeks full dismissal of Count VI (Unfair Competition by Misappropriation) on grounds of Copyright Act preemption. Infra Section IV(D). Continue Reading Here. Continue Reading Here . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case retrieved on February 26, 2024, from fingfx.thomsonreuters.com is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. fingfx.thomsonreuters.com

Stories, News & Mentions about THOMSONREUTERSCORP

Thomsonreuterscorp’s Stories on HackerNoon

Thomsonreuterscorp's News & Mentions