Meta Platforms is heading into its May 27, 2026 annual meeting on the heels of two jury losses that should change how investors think about the Company’s artificial intelligence ambitions.
National Legal and Policy Center is circulating an exempt solicitation report — and a full PDF version — to Meta investors urging a vote FOR Proposal Three on the 2026 proxy ballot. The proposal, sponsored by NLPC, asks Meta to publish an annual report assessing the operational, financial, and public-welfare risks of unethical or improper external data usage in its AI products. Shareholders do not need to wait for the annual meeting. They can cast their proxy votes today.
The case has gotten sharper since NLPC filed the proposal.
On March 24, 2026, a New Mexico jury ordered Meta to pay $375 million in civil penalties after finding the Company misled users and enabled child sexual exploitation on Facebook, Instagram, and WhatsApp. One day later, a Los Angeles jury found Meta and Google negligent for designing social-media products harmful to minors and awarded $6 million in damages.
Reuters described the back-to-back rulings as a fight over “tech‘s liability shield.” Both cases proceeded on a design-based theory of liability that narrows Section 230’s practical protection — and they arrived as Meta is racing to embed AI deeper into the same products at issue in those trials.
That race is not cheap. Meta has guided 2026 capital expenditures of $115–135 billion, nearly double 2025 levels, with much of it directed at AI infrastructure. In April, Meta Superintelligence Labs released Muse Spark, the first model in a new proprietary series the Company markets as “personal superintelligence” rooted in “the relationships and context already at the center of your life.”
Translation: Meta intends to build AI products fueled by the most intimate corners of users’ digital lives. And shareholders still have no board-level assessment of how the data feeding those products is sourced, vetted, or governed.
That gap is glaring against Meta’s governance record. The Company paid a record €1.2 billion GDPR fine for illegally transferring EU user data to the United States and settled the Cambridge Analytica litigation for $725 million. Meta then rewrote its privacy policy so an expansive range of user-generated images, videos, and text may be incorporated into AI training without users’ explicit, opt-in consent.
Meta’s Board urges a vote against the Proposal, claiming existing privacy practices, disclosures, and committee oversight are enough. The Board’s own opposition statement (see Page 67 at the link) tells investors otherwise. Meta admits that private chat content can be used to train its AI when “the user or someone in the chat chooses to share those messages” — a structural consent loophole that sweeps every other participant’s words, photos, and context into the training pipeline without their individual consent.
That is precisely the kind of practice a serious board-level risk assessment would surface. It is also precisely what Meta’s existing privacy disclosures do not capture.
Investors do not get to vote on whether AI eats Meta’s products. They do get to vote on whether management is candid about the risks. Proposal Three is a low-cost, high-information request: an annual report, omitting proprietary information, that tells shareholders what Meta knows about the legal, regulatory, reputational, and public-welfare exposure of its AI-data strategy.
The May 27 annual meeting is just over four weeks away. Meta investors should not wait for the gavel. Vote FOR Proposal Three now at www.proxyvote.com (you’ll need the control number or account number from your proxy card or notice).
NLPC’s executive summary is the place to start. The full report makes the rest of the case.
(Post references PX14A6G Notice of exempt solicitation)
