Meta Shareholders Warn of AI-Driven Disinformation: Here’s What You Need to Know

Meta is preparing for a shareholder showdown at its annual investor meeting today over the growing threats of generative artificial intelligence.

The tech giant’s shareholders are pushing it to share an annual report of the costs associated with gen AI—including financial, reputational, legal and regulatory—plus how it will protect people from misinformation and measure that effectiveness.

“With the issues with AI, we’re concerned with Meta’s inability to evaluate, particularly advertising and paid content related to the elections,” said Christina O’Connell, manager of shareholder engagement at Ekō, a corporate accountability organization representing Meta’s retail shareholders.  

Activist shareholder initiatives are common, but proposals relevant to AI are a more recent occurrence, according to Gary Kibel, a privacy and data security lawyer at Davis+Gilbert. Microsoft and Apple faced similar proposals at their respective annual meetings.

The upcoming U.S. and global elections—plus Meta’s plans to scale investment in AI from $5 billion to $40 billion by the end of 2024, per its first-quarter earnings report—make the matter more urgent for the company’s shareholders.

The proposal was co-filed by Arjuna Capital (asset manager), Open MIC and Ekō—all representing Meta’s investors.

What does Meta say?

Before the annual meeting, shareholders get to submit questions on resolutions filed with the company months in advance. The company then responds to these resolutions.

Perhaps unsurprisingly, Meta already opposes these AI resolutions, saying that its existing policies and self-regulatory measures are sufficient.

In its statement, the platform said its adoption of responsible AI principles, board oversight, investments in combating misinformation and disinformation and the development of watermarking tools provides “visibility into the impact of our products.”

Investors point out that none of these measures assess the return on that investment or the potential losses the company could face if it is found liable for harms associated with these unpredictable tools.

What’s the deal with Meta’s voting?

Shareholders will vote on the issue at the meeting.

“A very small number of people, primarily (Meta CEO) Mark Zuckerberg and a close circle of his associates, have special shares, and each one of their votes counts as 10 votes,” said O’Connell, “which is outrageous and not very democratic.”

As a result, each vote will be considered as an independent vote, O’Connell added.

What’s been the outcome of the other votes?

At its annual meeting last December, about 21% of Microsoft investors voted in favor of the AI-related shareholder proposal filed by Arjuna Capital, which asked the company to report on AI risks and draw out its contingency plans to limit harm.

Following the voting, Microsoft issued its first report. “It’s not everything we asked for, but it’s a good first step,” O’Connell said, but she would not share specifics.  

During Apple’s annual meeting in February, 37.5% of investors backed a shareholder proposal urging the company to disclose additional details regarding its utilization of AI and its ethical considerations.

Sounds like progress, right?

Well, sort of—there’s still a lot left to be desired. And with global elections right around the corner, the stakes are higher than ever.

Meta has already been caught greenlighting a series of AI-generated political ads running during India’s election campaign.

The tech giant—which laid off its trust and safety team last year—approved 14 ads calling for the killing of Muslims and spreading misinformation, according to an Ekō report this month shared with ADWEEK, first seen by The Guardian.

Meta acknowledged that these political ads incited violence in India.

What’s different about this vote?

Investors aren’t backing down, especially with elections looming large.

Today’s investor proposal also focuses on Meta’s content moderation efforts across five countries outside of the U.S., including India, Brazil and the Philippines.

“Meta, Alphabet and Microsoft are rushing to develop and deploy AI. They claim to have very strong policies about the use of AI, but they do not have the staff, mechanisms or programs to moderate and evaluate the threats from [AI],” said O’Connell. “That’s certainly a serious concern.”

Will Meta oblige?

Probably not.

“Shareholder initiatives are hard to get through since voting is controlled by institutional investors,” said Kibel. “They will not want to impose anything on the company that gets in the way of being profitable.”

Who’s next?

Alphabet faces a similar investor proposal on its AI efforts by shareholders June 7.

Ultimately, investors want assurances that tech giants are taking the necessary measures to safeguard platforms and people from the risk of AI.

“As long-term shareholders, we want Meta and Alphabet to succeed over the long run,” Natasha Lamb, chief investment officer at Arjuna Capital, said in a statement, “which means our companies must do what they can today to mitigate the gen AI risks of tomorrow.”