Gusinsky Trust v. Nadella Says Microsoft Board Hid AI Copyright Risks From Shareholders
Copyright and Biometric Violations Alleged in AI Training Data
The Vladimir Gusinsky Revocable Trust, suing derivatively on behalf of Microsoft Corporation, alleges that Microsoft’s board and executives caused the company to violate federal copyright and biometric privacy laws by training its AI models on pirated books and voiceprints without consent. The complaint, filed August 10, 2026, in the U.S. District Court for the Western District of Washington, claims Microsoft copied “hundreds of thousands of copyrighted books” to train its GPT models, despite public representations that its AI training data was obtained through “negotiated arrangements with publishers and copyright owners.” The complaint alleges, “Microsoft and OpenAI copied hundreds of thousands of copyrighted books without permission to train GPT models” (Compl. ¶3). Specifically, the trust cites allegations that Microsoft copied approximately 200,000 pirated books (Compl. ¶48).
The complaint further alleges that Microsoft’s commercial voice AI products extract voiceprints and biometric information at scale during base-model training. Microsoft’s foundational voice models, including Azure neural voices, Custom Neural Voice, Personal Voice, Voice Live API, and MAI-Voice-1, were trained on voiceprints extracted without speaker knowledge or consent. The trust cites two pending lawsuits—Basich v. Microsoft Corp. (2:26-cv-00422) and Flowers v. Microsoft Corp. (1:26-cv-05491)—as evidence of Microsoft’s alleged unlawful practices under the Illinois Biometric Information Privacy Act (BIPA). The complaint alleges that Microsoft ingested hundreds of thousands of hours of human speech recordings to build its voice AI products, which were then monetized through Azure, Teams, and Copilot (Compl. ¶53).
Microsoft’s VALL-E model was trained on 60,000 hours of data from over 7,000 speakers, while its NaturalSpeech 2 and 3 models were trained on 44,000 and 200,000 hours of speech data, respectively—all without speaker-by-speaker consent (Compl. ¶59). The complaint alleges that Microsoft extracted voiceprints from Teams meeting participants, journalists, podcasters, audiobook narrators, and voice actors, then commercialized these biometric signatures. As the complaint states, “Biometric Data” expressly includes “personal data resulting from specific technical processing relating to physical, physiological or behavioral characteristics of a natural person” (Compl. ¶56). The complaint also notes that Microsoft’s commercial voice AI products include Azure neural voices, Custom Neural Voice, Personal Voice, Voice Live API, MAI-Voice-1, Speaker Recognition API, Teams voice attribution, and research models such as VALL-E and NaturalSpeech (Compl. ¶54). These products, the complaint alleges, extract and process "biometric voice signatures" from audio recordings for speaker verification under GDPR Article 4 (Compl. ¶56). The complaint further alleges that Microsoft’s safeguards focused on downstream misuse of voice AI products rather than the upstream extraction of voiceprints without consent, stating, “If you are using [these products], you must ensure that you have the necessary rights, permissions, and licenses to use the voice data for training and deployment” (Compl. ¶57).
In its 2024 Proxy Statement, Microsoft represented that its AI training data was sourced from “negotiated arrangements with publishers and copyright owners” or synthetic datasets, and that it respected robots.txt and NO ARCHIVE tags. The complaint quotes Microsoft’s statement: “Generative AI Models — Microsoft uses a variety of data sources, including publicly available information, in a manner consistent with global copyright laws” (Compl. ¶2). The complaint further cites Microsoft’s claim that “We train on data that we gain access to... through negotiated arrangements with publishers and copyright owners” (Compl. ¶2). However, the complaint alleges these statements were materially false and misleading, as Microsoft’s AI models were trained on pirated books and voice data obtained without permission. The complaint alleges that Microsoft concealed these practices, stating, “Defendants concealed material problems with Copilot, including copyright violations, BIPA violations, and AI model underperformance” (Compl. ¶12). The complaint also alleges that Microsoft had violated federal copyright laws in the formulation of its AI strategy and its partnership with AI companies like OpenAI, as stated in the complaint: “Microsoft had violated federal copyright laws in the formulation of its AI strategy and its partnership with AI companies like OpenAI” (Compl. ¶12).
Board Oversight of OpenAI Partnership Allegedly Enabled Violations
The complaint alleges that Microsoft’s board of directors had direct oversight of Microsoft’s investment in OpenAI, which the complaint references as both $13 billion and $10 billion in different contexts (Compl. ¶43, 51). Under the partnership, OpenAI committed to purchasing $250 billion in incremental Azure services, while Microsoft committed to providing the computational infrastructure for OpenAI’s AI training (Compl. ¶68, 101). The board, according to the complaint, maintains direct oversight of AI strategy risk, with the Environmental, Social, and Public Policy Committee assisting in overseeing AI governance, regulatory risk, and responsible AI policies (Compl. ¶88).
In addition to its partnership with OpenAI, Microsoft announced a strategic partnership with Anthropic in November 2025, which included a $5 billion investment and a commitment by Anthropic to purchase $30 billion in Azure compute capacity (Compl. ¶69). Anthropic also committed to purchasing an additional 1 gigawatt of capacity from Microsoft (Compl. ¶69). The complaint alleges that these investments were overseen by the board and that the Director Defendants failed to address the legal risks associated with Microsoft’s AI training practices.
The complaint alleges that the board ignored a June 23, 2025, ruling by Judge William Alsup in Bartz v. Anthropic (3:24-cv-05417-WHA), which distinguished between lawful and pirated training data under the fair use doctrine. The ruling held that “Training on lawfully acquired books was ‘exceedingly transformative’ and could constitute fair use under 17 U.S.C. § 107,” but the complaint alleges Microsoft continued to use pirated data post-ruling, stating, “Downloading and [using] pirated books was not fair use” (Compl. ¶139). The complaint further alleges that the board consciously disregarded red flags, including the ruling and pending lawsuits, and failed to disclose these risks to shareholders. As the complaint states, “Defendants breached duties of care, loyalty, and good faith by causing Microsoft to make false/misleading statements about AI strategy, Copilot products, and financial results; violated copyright and biometric privacy laws; and concealed material problems with Copilot” (Compl. ¶1, FIRST CAUSE OF ACTION). The complaint also alleges that the Director Defendants disseminated materially false and misleading statements in Microsoft’s 2024 and 2025 Proxy Statements, inducing shareholders to re-elect them and approve executive compensation (Compl. SECOND CAUSE OF ACTION).
Microsoft’s board, according to the complaint, approved the design and deployment of bespoke Azure supercomputing infrastructure optimized for training AI models on large-scale datasets, including copyrighted and biometric data. The complaint alleges that the board knew or recklessly disregarded that Microsoft’s AI training practices violated copyright and biometric privacy laws, yet failed to disclose these risks in its SEC filings. The complaint brings a claim for breach of fiduciary duty against all Individual Defendants, including the Director Defendants, for their alleged failure to oversee these practices (Compl. FIRST CAUSE OF ACTION). The complaint alleges that the Director Defendants were directly involved in the supervision of the $13 billion OpenAI partnership, which the complaint alleges caused copyright and BIPA violations (Compl. ¶136). The complaint further alleges that the board approved false and misleading public statements about Microsoft’s AI products, including statements in the company’s proxy statements and earnings calls, and that the board’s failures included the approval of stock repurchases while these risks were concealed (Compl. ¶145-148).
Proxy Statements and Earnings Calls Allegedly Misrepresented AI Adoption and Performance
The complaint alleges that Microsoft’s 2024 and 2025 Proxy Statements contained materially false and misleading statements about the performance and legal compliance of its AI products, including Microsoft 365 Copilot and Azure AI. The trust brings a claim under Section 14(a) of the Securities Exchange Act, alleging that the Director Defendants disseminated false and misleading statements in its proxy materials to induce shareholders to re-elect directors and approve executive compensation (Compl. SECOND CAUSE OF ACTION). The complaint alleges that the Director Defendants “disseminated materially false and misleading statements in Microsoft’s 2024 and 2025 Proxy Statements, inducing shareholders to re-elect them and approve executive compensation” (Compl. SECOND CAUSE OF ACTION).
In its 2025 Proxy Statement, Microsoft claimed that 90% of Fortune 500 companies were using Microsoft 365 Copilot, with 230,000 organizations using Copilot Studio and 70,000 customers using Azure AI Foundry (Compl. ¶80, 86, 91). The complaint alleges these figures were misleading, as later reports revealed that only 15 million paid Copilot seats had been sold as of Q2 2026, out of 430 million M365 Commercial paid seats (Compl. ¶81, 112). The complaint cites a February 2026 Wall Street Journal report describing Copilot’s “confusing brand positioning and interoperability problems” and noting that only a small proportion of subscribers to Microsoft’s enterprise suite used the product. The complaint quotes the Journal: “Confusing brand positioning and interoperability problems have frustrated users” and “Only a small proportion of subscribers to Microsoft’s enterprise suite use Copilot” (Compl. ¶114). The complaint further alleges that Microsoft violated Item 303 of SEC Regulation S-K by failing to disclose known trends and uncertainties related to Copilot’s performance, adoption, and legal risks in its Management’s Discussion and Analysis of Financial Condition and Results of Operations. Specifically, the complaint alleges that Microsoft failed to disclose that its AI strategy exposed the company to significant legal risks, including copyright and BIPA violations (Compl. VIOLATION OF ITEM 303). The complaint also alleges that Microsoft violated Item 105 of SEC Regulation S-K by providing inadequate disclosure of risks associated with its AI strategy, including the legal risks of its training practices and the competitive challenges facing Copilot (Compl. VIOLATION OF ITEM 105).
During Microsoft’s Q3 2025 earnings call on April 30, 2025, CEO Satya Nadella and CFO Amy Hood highlighted the company’s AI-driven growth, with Nadella stating that Microsoft was investing “for any workload” in AI models, whether open or closed source (Compl. ¶77). The company reported Intelligent Cloud revenue of $26.8 billion for the quarter, a 21% increase year-over-year, with Azure and cloud services revenue growing 33% (Compl. ¶78). For fiscal year 2025, Microsoft reported Intelligent Cloud revenue of $29.9 billion in Q4, a 26% increase, with Azure and cloud services revenue growing 39% (Compl. ¶82). The complaint alleges these statements were materially misleading, as they failed to disclose the legal and operational challenges facing Microsoft’s AI products. The complaint also notes that Microsoft’s capital expenditures for fiscal year 2025 totaled $88 billion, with the company adding 2 gigawatts of capacity (Compl. ¶93).
By Q1 2026, Microsoft reported Intelligent Cloud revenue of $30.9 billion, a 28% increase, with Azure and other cloud services revenue increasing 40% (Compl. ¶99). However, the complaint alleges that these figures masked underlying problems, including the diversion of computational capacity from Azure to Copilot and AI research and development. The complaint cites a February 11, 2026, report by Melius Research linking Copilot’s struggles to Azure capacity constraints, stating, “Copilot Woes and Azure are Linked” (Compl. ¶116). The complaint also highlights that Microsoft’s capital expenditures for the first six months of fiscal 2026 reached $72.4 billion, reflecting the company’s significant investment in AI infrastructure (Compl. ¶119). The complaint further alleges that Microsoft’s 2025 Proxy Statement touted Nadella’s leadership, attributing fiscal 2025 results to AI-driven innovation, security, and quality, and justifying his 151.67% performance payout rating as due to his “thoughtful leadership” (Compl. ¶89, 95). The Proxy Statement also represented that Microsoft’s AI training complied with global copyright laws and excluded data from the USTR Notorious Markets list, claims the complaint alleges were false and misleading (Compl. ¶90).
Stock Drop and Insider Sales Allegedly Followed Undisclosed AI Problems
The complaint alleges that Microsoft’s stock price was artificially inflated by the company’s false and misleading statements about its AI products, and that the stock suffered a significant decline once the truth about Copilot’s struggles was disclosed. On July 31, 2025, Microsoft stock reached an intraday all-time high of $551.05 per share (Compl. ¶07/31/2025). However, on January 28, 2026, Microsoft’s stock closed at $481.63 per share, but the next day, after the company announced its Q2 2026 results, the stock fell to $433.50 per share—a $48 per share drop on unusually high trading volume of 129 million shares (Compl. ¶113). The complaint alleges that this decline was triggered by the disclosure of Copilot’s “perception” issues, which the board had allegedly known about but failed to disclose. As CFO Kathleen McBride acknowledged in June 2026, “The product is way different than it was 90 days ago” (Compl. ¶120).
The complaint alleges that Nadella sold Microsoft stock at inflated prices before the stock drop. The complaint states that Nadella sold over $75 million in stock at prices exceeding $500 per share, compared to post-disclosure lows (Compl. ¶129). Nadella’s total compensation for 2023–2025 was $224,115,510, including $96,496,790 in 2025 alone, which the complaint alleges was unjust due to his alleged breaches of fiduciary duty (Compl. ¶150). The complaint details Nadella’s compensation breakdown for 2025 as follows: $2,500,000 in salary, $84,245,496 in stock awards, $9,555,000 in incentive compensation, and $196,294 in other compensation (Compl. ¶150). For 2024, Nadella received $79,106,183 in total compensation, including $2,500,000 in salary, $71,236,392 in stock awards, $5,200,000 in incentive compensation, and $169,791 in other compensation (Compl. ¶150). For 2023, his total compensation was $48,512,537, comprising $2,500,000 in salary, $39,236,137 in stock awards, $6,414,750 in incentive compensation, and $361,650 in other compensation (Compl. ¶150). The complaint notes that Nadella’s 2025 compensation was 480 times the median Microsoft employee’s total compensation of $200,972 (Compl. ¶151). The complaint brings a claim for unjust enrichment against the Executive Officer Defendants, including Nadella, Hood, Spataro, and Jha, alleging that they “were unjustly enriched by receiving excessive compensation while breaching fiduciary duties and causing Microsoft to violate copyright and biometric privacy laws” (Compl. THIRD CAUSE OF ACTION).
In addition to Nadella, the complaint details the compensation of other executives, including CFO Amy Hood, who received $25,799,206 in total compensation for 2024 and $19,902,897 for 2023 (Compl. ¶150). For 2025, Hood’s total compensation was $29,481,551, including $25,037,360 in stock awards, $1,000,000 in salary, $3,421,000 in non-equity incentive plan compensation, and $23,191 in other compensation (Compl. ¶150). The complaint alleges that these compensation packages were unjust given the executives’ alleged breaches of fiduciary duty and the company’s legal and operational challenges. The complaint further alleges that the Executive Officer Defendants were unjustly enriched by receiving excessive compensation while the company faced significant legal and financial risks, stating, “Executive Officer Defendants were unjustly enriched by receiving excessive compensation while breaching fiduciary duties and causing Microsoft to violate copyright and biometric privacy laws” (Compl. THIRD CAUSE OF ACTION).
The complaint alleges that Microsoft’s capital expenditures for fiscal year 2026 were projected to reach $190 billion, $37 billion above analyst expectations, due in part to the diversion of computational capacity from Azure to Copilot and AI research and development (Compl. ¶119). The complaint cites Microsoft’s announcement of $190 billion in planned capital expenditures for fiscal year 2026, which included $37.5 billion for Q2 2026 alone (Compl. ¶119). The complaint alleges that these expenditures were driven by the need to support Copilot and AI development, which diverted resources from Azure and other core businesses. The complaint further alleges that Microsoft’s Azure growth slowed due to CPU and GPU capacity being diverted to Copilot and AI R&D, with only 15 million paid Microsoft 365 Copilot seats sold as of Q2 2026, despite 450 million commercial users (Compl. ¶112). The complaint also highlights that disorganized data silos were an issue for Copilot, as noted by Citi Research: “Disorganized data silos have been an issue for Copilot” (Compl. ¶128).
The complaint also highlights the competitive challenges facing Copilot, noting that only 11.5% of enterprise users preferred it as of February 2026, down from 18.8% (Compl. ¶114). The complaint quotes Microsoft EVP Rajesh Jha’s statements during the Q1 2026 earnings call, where he claimed, “two quarters in a row, the daily active engagement with Copilot has more than doubled, quarter over quarter” and “what, 70% of the Fortune 500 had Copilot. Now, that’s up to 90%” (Compl. ¶103). The complaint alleges these statements were misleading, as they did not reflect the actual adoption or performance of Copilot. Jha also described Copilot’s “Work IQ” as providing superior work context understanding compared to competitors’ connectors and asserted that Copilot uniquely integrates AI into existing workflows, such as Excel and PowerPoint (Compl. ¶105). The complaint quotes Jha: “the reasoning model finally unlocks AI and graphical user interfaces to work much more effectively” (Compl. ¶104).
Demand Futility Alleged Under Delaware Law
The complaint alleges that demand on Microsoft’s board to pursue this litigation would have been futile under Delaware law, as a majority of the directors face a substantial likelihood of liability for their alleged breaches of fiduciary duty. The complaint alleges that the board consciously ignored red flags of large-scale copyright and biometric privacy violations, including the 2025 federal ruling in Bartz v. Anthropic and the pending lawsuits against Microsoft. The complaint cites the board’s alleged failure to address these risks in its oversight of Microsoft’s AI strategy, stating, “Director Defendants disseminated materially false and misleading statements in Microsoft’s 2024 and 2025 Proxy Statements, inducing shareholders to re-elect them and approve executive compensation” (Compl. SECOND CAUSE OF ACTION). The complaint further alleges that the board’s oversight failures violated Item 303 of SEC Regulation S-K by failing to disclose known trends and uncertainties related to Copilot’s performance, adoption, and legal risks, as well as Item 105 by inadequately disclosing the risks associated with Microsoft’s AI strategy, including copyright and BIPA violations (Compl. VIOLATION OF ITEM 303, VIOLATION OF ITEM 105).
The complaint alleges that the board’s oversight failures extended to its Audit Committee, which included directors Hugh Johnston, Teri List, and John Stanton, as well as its Environmental, Social, and Public Policy Committee, which included Penny Pritzker, Reid Hoffman, John Stanton, Emma Walmsley, and Catherine MacGregor (Compl. ¶153, 157). The complaint alleges that these committees failed to address the legal and operational risks associated with Microsoft’s AI strategy, despite their stated responsibilities. For example, the 2024 Proxy Statement claimed that the Board oversees AI strategy risk via the Environmental, Social, and Public Policy Committee, but the complaint alleges that the committee failed to address the unlawful training practices and legal risks (Compl. ¶72). The complaint also notes that the 2024 Proxy Statement included shareholder proposals raising concerns about AI risks, including Proposal No. 9, which alleged that OpenAI scraped private data without permission and cited the New York Times copyright lawsuit (Compl. ¶74). Despite these concerns, the Director Defendants recommended voting against the proposals, claiming existing "Responsible AI" frameworks were sufficient (Compl. ¶75). The complaint quotes the Proxy Statement’s representation that Microsoft trained its AI on data from “negotiated arrangements” with publishers and copyright owners or synthetic datasets, and that it respected robots.txt and NO ARCHIVE tags, allowing opt-outs for AI training (Compl. ¶75).
The complaint alleges that the board’s failures were not limited to oversight but also included direct participation in the supervision of Microsoft’s $13 billion OpenAI partnership, which the complaint alleges caused copyright and BIPA violations (Compl. ¶136). The complaint further alleges that the board approved false and misleading public statements about Microsoft’s AI products, including statements in the company’s proxy statements and earnings calls (Compl. ¶145-148). The complaint seeks damages, disgorgement of unjust compensation, and corporate governance reforms, including changes to the board’s oversight of AI strategy and risk management. The complaint alleges that the board’s failures included the approval of Microsoft’s 2025 Proxy Statement, which touted the company’s AI leadership under Nadella and attributed fiscal 2025 results to AI-driven innovation, security, and quality (Compl. ¶89). The Proxy Statement also represented that Microsoft’s AI training complied with global copyright laws and excluded data from the USTR Notorious Markets list, claims the complaint alleges were false and misleading (Compl. ¶90). The complaint further alleges that the board maintained direct oversight over AI strategy risk and that the Environmental, Social, and Public Policy Committee assisted in overseeing AI governance, regulation risk, and responsible AI policies (Compl. ¶88).
The complaint also details the board’s alleged failure to disclose material risks in its SEC filings, including the company’s Form 10-Q for Q1 2026, which reported Intelligent Cloud revenue of $30.9 billion and Azure revenue growth of 40% (Compl. ¶99). The complaint alleges that these filings violated Item 303 of SEC Regulation S-K by failing to disclose known trends and uncertainties related to Copilot’s performance and legal risks, as well as Item 105 by inadequately disclosing the risks associated with Microsoft’s AI strategy (Compl. VIOLATION OF ITEM 303, VIOLATION OF ITEM 105). The complaint further alleges that the board failed to address the diversion of computational capacity from Azure to Copilot, which increased Microsoft’s capital expenditures by billions of dollars. The complaint cites Microsoft’s $72.4 billion in capital expenditures for the first six months of fiscal 2026 and its planned $190 billion in capital expenditures for the full fiscal year, $37 billion above analyst expectations (Compl. ¶119). The complaint also alleges that the board’s failures included the approval of Nadella’s $224 million+ compensation for 2023–2025, which the complaint alleges was unjust due to his fiduciary breaches (Compl. ¶150).
The allegations in the complaint are unproven, and no defendant has yet responded to the claims. Microsoft has not commented on the litigation.