Technology-Related Proposals in the AI Era: Notes from the 2026 Proxy Season

Growing concerns among investors, regulators, companies, consumers, and other stakeholders about the economic, environmental, and social impacts of artificial intelligence (AI) and data centers have contributed to the recent prominence of shareholder proposals covering these issues on proxy ballots. Since they were first proposed in 2024, 29 AI-related proposals have gone to a vote, including 11 in 2026.^1,2 Although the number of AI-related proposals has remained relatively stable year-to-year, the 2026 proxy season highlighted the breadth of issues that investors increasingly associate with AI and data center expansion. This article covers some notable examples of AI-related shareholder proposals that touch on consumer data privacy, AI data sourcing, energy demand, water use and electricity costs.

Expanding Investor Concerns: Data Privacy and Sourcing

Alphabet's 2026 annual meeting provides a useful illustration of this expanding range of investor concerns. The data privacy and AI data sourcing proposals presented at this meeting highlighted how AI has broadened traditional concerns over consumer-data protection by introducing a wider set of legal, reputational, and operational risks.³

Data privacy proposals have focused primarily on consumer impacts arising from gaps in corporate policies, practices, and oversight that result in the misuse or mishandling of consumer data. The risks associated with data privacy are relatively well-defined, and include legal and regulatory challenges arising from the EU’s General Data Protection Regulation (GDPR) enforcement actions as well as litigation and regulatory scrutiny related to the California Consumer Privacy Act and other privacy-related legislation. For example, Zevin Asset Management's data privacy proposal at Alphabet raised concerns over (i) consent for targeted advertising data; (ii) tracking user data from “private” internet searches; (iii) misusing cell phone data; and (iv) inappropriately accessing user data with AI products.⁴ The proposal received 6% support.

Whereas data privacy proposals generally address well-established risks, data sourcing proposals focus on how companies collect data to train AI products, raising concerns about a less defined and evolving set of risks. At Alphabet, the AI data sourcing proposal submitted by the National Legal and Policy Center (NLPC) reflected potential exposure arising from the improper or unethical use of data sourced from publicly available content, including copyrighted works such as art, literature, music, and journalism.⁵ It received 12.3% support, comparable to the 10.2% support received by a similar NLPC proposal presented at Meta Platforms’ annual meeting this year.

Beyond sourcing risks, customer use of AI raises issues regarding accuracy, misinformation and disinformation, biases, defamation, and a range of liabilities that regulators are now considering. A proposal at Alphabet submitted by Vancity Investment Management sought to address these concerns by requesting an independent assessment of additional actions the company could take to mitigate AI-generated misinformation.⁶ The need to account for customer behavior can require a more complex risk-management process than applies when addressing traditional data privacy concerns, a challenge that Alphabet has attempted to manage in part by stating that it would assume responsibility for potential legal risks if customers were challenged on copyright grounds. The proposal received 9.3% support.

Voting support levels for these proposals suggest that standards and expectations for data-related risk management remain nascent, with many investors opting to defer to the board on relevant disclosure and oversight. That said, both Alphabet and Meta have dual or multi-class voting structures that serve to understate support among minority shareholders; calculated on a one-vote-per-share basis, the data sourcing proposals at Alphabet and Meta were among the most highly supported AI-related resolutions going to a vote this season.

Climate and Environmental Risks as Material

Environmental risks associated with AI-driven data center expansion have also emerged as a prominent focus of shareholder proposals. Similar to the prior proxy season, Amazon.com faced a proposal in 2026 concerning the alignment of its climate commitments and its planned AI data center expansion, with comparable proposals also submitted at Alphabet and Meta.⁷ The Amazon proposal received notable support, albeit slightly diminished from 2025, dropping from 20.1% to 18.4%. The results underscore the relevance of climate and energy-related risks associated with data center expansion, particularly as Amazon's web services business now accounts for over half of its operating income. Shareholder support was notably lower at Meta and Alphabet, with 6.9% and 7.4% of shareholders voting in favor, respectively, though these figures should be interpreted in the context of the companies' voting structures.

Meanwhile, investors filed proposals at Alphabet and Digital Realty Trust to address water-use risks associated with AI-driven data center expansion. The proposal at Digital Realty, the world's largest data center REIT, requested region-level disclosure of water-risk and water-intensity data⁸ and received moderate support from 22.5% of voting shareholders. By contrast, the proposal at Alphabet⁹ sought a report on how the company's water-usage policies and practices align with its fiduciary duty to maximize long-term financial performance and received only 1.5% support. The stronger support for the Digital Realty proposal may reflect both its exposure to data center-related water risks and the proposal's focus on decision-useful information regarding regional water stress, operational water intensity, and exposure to water-scarcity constraints. In contrast, the proposal at Alphabet did not specify comparable metrics or reporting parameters, limiting its usefulness for shareholder analysis.

As the environmental and social impacts of AI-driven data center expansion attract greater scrutiny, shareholder proponents may increasingly concentrate on specific concerns related to power consumption, water and land use, grid reliability, and electricity costs. For example, a proposal at The Southern Company in 2026, which received 9.7% support, examined for the first time the effect of data centers on residential and small-business electricity bills and questioned whether the company had safeguards in place to prevent cost shifting.¹⁰

While companies’ investments in and use of AI technologies present significant growth opportunities, there are also significant risks, including a variety of legal, regulatory, and reputational risks. As such, companies should ensure appropriate oversight and risk mitigation strategies are in place to protect and/or enhance shareholder value. Nevertheless, shareholders appear ready to afford boards significant discretion in determining how oversight responsibilities should be structured, providing that companies can demonstrate that material risks are being effectively managed.