Resolved clauseShareholders request the Company prepare a report, at reasonable cost, omitting proprietary or legally privileged information, to be published within one year of the Annual Meeting and updated annually thereafter, which assesses the risks to the Company’s operations and finances, and to public welfare, presented by the real or potential unethical or improper usage of external data in the development and training of its artificial intelligence offerings; what steps the Company takes to mitigate those risks; and how it measures the effectiveness of such efforts.
Whereas clauseThe immense and transformative potential of artificial intelligence comes with substantial risks.
The development and training of AI systems rely on vast amounts of data, and public information available via the Internet may not be enough to quench developers’ insatiable thirst for high-quality training data.1,2 Thus, stakeholders are concerned that developers will draw from unethical or illegal sources – such as personal information collected online,3 copyrighted works,4 and proprietary commercial information provided by users.5
Supporting statementMicrosoft Corporation (“Microsoft” or the “Company”) is an early leader in the AI arms race6,7 because of its extensive partnership with OpenAI8 and early dominance in business-to-business solutions, 9 which has helped push the Company to the highest market capitalization in the world.10
But shareholders should be concerned with Microsoft’s record on data ethics:
• Microsoft employs generative AI models developed by OpenAI, which allegedly stole large amounts of personal information by scraping the web, including “private information and private conversations, medical data, information about children – essentially every piece of data exchanged on the internet it could take – without notice to the owners or users of such data, much less with anyone’s permission.”11
• Paul Nakasone, former director of National Security Agency – which has been criticized for spying on American citizens – sits on OpenAI’s board.12,13
• Microsoft received pushback against its proposed AI “Recall” feature, which screenshots everything a Windows and Copilot user sees or does and turns it into searchable data. The feature was delayed due to the resistance, but then was eventually released.14,15,16,17
• Microsoft or OpenAI, through their services, may inadvertently or deliberately access and utilize proprietary information provided by users.18,19
• Microsoft and OpenAI have been sued by the New York Times, among others, which alleged copyright infringement.20
Microsoft is now developing its own models to better compete, including against OpenAI.21 The Company has been light on disclosures regarding what user data may be retained by itself or its partners for training purposes. Further, the Company has no publicly available metrics to measure the success of its AI data ethics efforts.
Prioritizing data ethics in Microsoft’s AI development may reduce regulatory risk22,23,24 and provide a competitive advantage with consumers.25 Microsoft’s position in the AI arms race, and its associated historic valuation, hang in the balance.