Next-Gen Governance: AI’s Role in Shareholder Proposals

With AI continuing to captivate businesses, consumers and regulators, institutional investors are also increasingly focusing their attention on the implications for the businesses in which they invest. In our Responsible AI Governance white paper published in January 2024, we highlighted the risks to businesses and wider society of AI, provided an overview of regulators’ and investors’ responses, and proposed a governance framework to manage these risks. One topic covered in the report that is gaining traction is the emergence of shareholder proposals urging companies to disclose more information on the risks associated with  the deployment of AI and their efforts to mitigate them.

An investment trust for union members, the American Federation of Labor and Congress of Industrial Organizations (AFL-CIO) filed an AGM proposal demanding greater transparency on AI at five entertainment companies (Apple, Comcast, Disney, Netflix and Warner Bros). It also filed a proposal at Amazon asking for a new committee of independent directors on AI to address human rights issues. Similarly, Arjuna Capital, an investment firm focused on sustainability, has filed a proposal at three technology companies (Alphabet, Meta and Microsoft) requiring them to issue a report on the risks of misinformation and disinformation resulting from AI. Alphabet also received a proposal from Trillium ESG Global Equity Fund requesting that the company clarify board committee responsibility and oversight for AI. SHARE, a Canadian responsible investment association, also filed a proposal at United Health Group calling for a report on the use and oversight of AI in operations. Two of these proposals have now been voted on (at the Microsoft AGM on 7 December 2023 and at the Apple AGM on 28 February 2024) and although they both failed it is interesting to evaluate them. Additionally, three of the companies (Comcast, Disney and United Health Group) did agree to disclose more AI-related information, leading AFL-CIO and SHARE to withdraw their bids.

This note focuses on the response of leading proxy advisors, Glass Lewis and ISS, to this new type of shareholder proposal with an emphasis on the one filed by AFL-CIO at Apple. We have  concentrated on this proposal since it was endorsed by both proxy advisors and their analyses encompassed similar and additional aspects compared to those reviewed in the context of Arjuna’s proposal the Microsoft AGM. As major institutional investors typically receive the research and vote recommendations produced by one or both of these advisors prior to voting at shareholder meetings, their stance carries significant weight, especially for a new and complex topic like AI.

At Apple, Glass Lewis and ISS supported AFL-CIO’s proposal arguing that improved transparency would allow shareholders to better evaluate the risks associated with the use of AI and would not be overly burdensome on the company. Notably, investors like Norges Bank Investment Management (NBIM) (who has publicly set out its expectations regarding the governance of AI by its portfolio companies), Legal & General Investment Management (LGIM) (who has also set out it expectations for AI adoption) and abrdn publicly expressed support for the proposal at the AGM. The resolution ultimately received 37.5% support.

At Microsoft, Glass Lewis and ISS opposed Arjuna’s proposal, as they felt Microsoft had already provided substantial disclosure on AI. Investors such as the Dutch asset manager APG and Californian public pension giant CalPERS also concluded that that support for this proposal was not warranted. However, several other large investors did support it including NBIM, the Office of the New York City Comptroller, which manages the city’s five pension funds, Californian public pension fund CalSTR, and Dutch pension fund manager PGGM. The proposal ultimately received 21.2% of the vote, representing a lower level of support than for the Apple proposal but still a significant minority of support.

Given the growing adoption of AI by businesses, it’s conceivable that we will observe more of these proposals in the future. We can also expect that AI will feature in the considerations of shareholders as they file proposals on other societal concerns such as labour, health and safety and the environment. We will continue to monitor these developments and review how proxy advisors and major investors approach such proposals. Moreover, we can also assist issuers in mitigating activism risks by supporting the development of responsible AI governance frameworks.

AFL-CIO Proposal at Apple Inc.

Text of the Proposal

The proposal filed by AFL-CIO at Apple stated that:

“Shareholders request that Apple Inc. prepare a transparency report on the company’s use of Artificial Intelligence (“AI”) in its business operations and disclose any ethical guidelines that the company has adopted regarding the company’s use of AI technology. This report shall be made publicly available to the company’s shareholders on the company’s website, be prepared at a reasonable cost, and omit any information that is proprietary, privileged, or violative of contractual obligations.”

AFL-CIO’s Perspective

AFL-CIO contended that implementing an ethical framework for AI adoption would enhance Apple’s standing “as a responsible and sustainable leader in its industry.” AFL-CIO noted that businesses’ use of AI raised important social concerns related to:

• Discrimination and bias against employees,
• The potential for mass layoffs,
• Violations to privacy of customers and members of the public, and
• Generating ‘deep fake’ content resulting in disinformation during elections.

The proposal recognised the ethical guidelines set by the White House Office of Science and Technology Policy for guiding the development, use and implementation of AI. Its five principles are:

1. Safe and effective systems,
2. Algorithmic discrimination protections,
3. Data privacy,
4. Notice and explanation, and
5. Human alternatives, consideration, and fallback.

AFL-CIO also emphasised that companies developing AI should refrain from using copyrighted materials or the attributes of professional performers “without transparency, consent and compensation to creators and rights holders.” Moreover, they opposed the use of AI to create literary content, advocating against its role in supplanting professional writers.

Apple’s Perspective

Apple’s board expressed reservations regarding the broad scope of the requested transparency report, explaining that it could expose strategic plans and initiatives important to maintaining the company’s competitive position.

Highlighting the company’s dedication to responsibly integrating AI into its products and services, the board highlighted existing transparency and resources. It also argued that the social issues raised by AFL-CIO were not unique to AI and were adequately addressed through current company guidelines and policies.

Moreover, the board deemed the proposal premature, citing the nascent regulatory landscape and the new legislative initiatives being developed across the world. It reiterated the board’s oversight of corporate and product strategy, with the regular updates it receives and the audit committee’s support in tracking important operational and reputation risks related to human right concerns.

Proxy Advisors’ Vote Recommendation & Rationale

In the below table we outline Glass Lewis’ and ISS’ voting recommendations on the Apple proposal and their rationale by highlighting the factors they analysed.

Further background on AFL-CIO

Glass Lewis highlighted that AFL-CIO submitted four shareholder proposals in the first half of 2023 and these resolutions received an average support of 29.4%. This indicates that AFL-CIO is an experienced filer who has proven its ability to generate traction for its causes. In 2023, some of AFL-CIO’s efforts focused on asking companies to adopt and disclose a policy on their commitments to respect the international human rights of freedom of association and collective bargaining.

FTI View

Glass Lewis’ and ISS’ evaluation of the AFL-CIO proposal at Apple align with their overarching approach to assessing ESG shareholder initiatives. They tend to be more inclined to support proposals that seek enhanced disclosures related to material risks, as is the case here. However, as observed in previous instances, such as with climate-related proposals, proxy advisors and shareholders are less inclined to support overly broad or prescriptive proposals that could lead to excessive interference in a company’s management.

In contrast, it is interesting to note the proposal Arjuna Capital submitted for the Microsoft December 2023 annual meeting. This proposal requested a report detailing the potential risks to the company’s operations, finances, and broader public welfare arising from misinformation and disinformation disseminated or generated through AI. Arjuna Capital also requested an outline of the measures undertaken by the company to address these issues and an assessment of their effectiveness.

While that proposal was also principally about enhancing disclosure, both Glass Lewis and ISS recommended that shareholders vote against it. Their rationale primarily stemmed from the belief that Microsoft had already provided substantial disclosures regarding its considerations related to AI and the measures it had implemented to mitigate associated risks.

Looking Forward

AFL-CIO put forward AI shareholder proposals to four additional entertainment companies (Amazon, Disney, Comcast, Netflix and Warner Bros) while Arjuna Capital filed similar proposals at two other technology companies (Alphabet and Meta). In each case, these filers appear to have replicated their model proposal at the different companies. Alphabet also received a proposal form Trillium and SHARE filed a proposal at United Health Group.

The last AGM where shareholders were expected to vote on an AI proposal was that of Disney. Disney, like Apple, engaged with the US Securities and Exchange Commission (SEC) in an attempt to remove the AFL-CIO’s AI proposal from the agenda of its 2024 AGM. However, the SEC denied both companies’ requests, concluding that “the proposal transcends ordinary business matters and does not seek to micromanage the company.” Despite this, Disney did not include AFL-CIO’s proposal in the meeting notice issued on 1 February 2024. It later emerged that AFL-CIO agreed to withdraw its bid following Disney’s commitment to disclose more information on its use of AI.  

AI is also emerging as a component of activist campaigns for board seats. In a parallel development at Disney, two activist investors, Blackwells Capital and Trian Fund Management, sought to appoint directors to Disney’s board at the 2024 AGM. Notably, Blackwells Capital’s arguments focused on a perceived need to strengthen Disney’s governance, transparency, content and technology strategy and their action plan included developing an AI strategy and disclosing it to shareholders. While the AFL-CIO proposals, seeking more disclosure on AI risks and risk mitigation strategies would generally be approached by proxy advisors like ESG shareholder proposals, Blackwells Capital’s proposals will be viewed through their traditional framework for contested director elections. That framework assesses the need for change and which nominees are best placed to create value for the company’s shareholders. None of the activists’ candidates were appointed at the AGM.

Considering the important opportunities and risks associated with AI, and the growing adoption of the technology across all industries, it is likely that a variety of shareholder proposals relating to AI will become increasingly prevalent. Investors are undeniably concerned and want to ensure management teams have the right expertise and risk mitigation processes in place to protect long term value. Proposals  will also increase as AI becomes another lever activists can use to apply pressure. We will continue to monitor the shareholder proposal landscape and the approaches of leading proxy advisors and investors.

One avenue for companies seeking to mitigate the risk of activism related to their AI practices is to adopt a proactive approach to the governance of AI such as the one outlined in our latest Responsible AI Governance report.