Krist Novoselic, a founding member and bassist of the renowned rock band Nirvana, recently took the helm in presenting a shareholder proposal, urging Microsoft to reevaluate its strategy in developing artificial intelligence (AI).
The proposal, titled "Shareholder Proposal 13: Report on Artificial Intelligence Misinformation and Disinformation," was submitted by Arjuna Capital on behalf of Krist Novoselic and several other shareholder groups, as indicated in a press release.
Novoselic, representing the Nirvana co-founder and band member, emphasized concerns that Microsoft might not have adequately assessed the business and societal risks associated with the rapid deployment of generative AI technology. The proposal voiced worries from significant shareholders about the potential for Microsoft-backed models to contribute to the dissemination of extensive disinformation and misinformation.
The proposal also queries whether Section 230, a law granting limited liability protection to internet hosts and website users for third-party content, extends to content generated by the host's own generative AI systems. Novoselic pondered the consequences of society relying on information produced by Microsoft's AI-powered Bing, a platform that reportedly produces inaccurate responses approximately 10% of the time. This proposal, which Novoselic initially suggested earlier this year, echoes prior calls by experts for a six-month moratorium on AI development, which were disregarded by Microsoft and other industry players.
Novoselic criticized Microsoft's haste in entering the market, alleging that it seems to prioritize short-term gains over long-term achievements. Although Microsoft's board of directors asserted that the proposal's requirements were met through existing and forthcoming reports, Arjuna Capital contends that the proposal seeks information beyond the general commitments to responsible AI outlined in these reports.
The shareholders behind the proposal aim to compel Microsoft to thoroughly evaluate the enduring risks linked with generative AI. Despite the board's recommendation to shareholders to dismiss the proposal based on existing plans and reports, the proposal failed to secure approval in a subsequent shareholder vote.
