Social Media When Elections Are Close: Doing the Impossible Better

Everyone knows, or should, that perfect (balanced, targeted, transparent, swift and swiftly appealable) content moderation at scale is impossible. That doesn’t mean powerful social platforms don’t have to try to get at least reasonably close, especially when access to good information can have a profound effect on political participation and speech.

That’s why so many people, including the Meta Oversight Board, criticized Meta for suspending a major political Instagram account for hours during a huge early voting push on October 17; and for allegedly boosting one candidate in Brazil’s recent presidential election. Meta denies deliberate interference in either instance, blaming “mistakes” in its automated content moderation systems (including AI tools that can’t tell the difference between enthusiastic follower engagement and a bot attack).

Mistakes or not, with early voting in the U.S. midterms already underway, these events serve as a stark reminder that all major social media platforms need to have stronger systems in place to ensure that they serve, rather than impede, democratic communication.

Access to good information can have a profound effect on political participation and speech.

This is not inherently an “AI” problem: automated systems can process content at a scale that humans never could, potentially enabling better moderation at scale and alleviating the psychological load on ill-paid moderators whose jobs require them to view incredibly disturbing content. But automated systems also reproduce existing biases, struggle to understand context, and often make mistakes that disproportionately affect journalists, activists, artists, and other vulnerable and marginalized communities. And moderation errors are even more likely when the language is spoken principally by the global majority. In 2020, leaked document showed that Meta’s algorithms incorrectly deleted nonviolent Arabic-language content 77 percent of the time, while failing to detect hate speech under the company’s own policies in many instances. Where these systems are faltering in Arabic, they’re positively failing in less-resourced languages. As a 2025 report from the Center for Democracy and Technology found, labeled datasets in certain languages and dialects such as Maghrebi Arabic and Kiswahili contain inconsistencies, bias, and inaccuracies due to the limited hiring of annotators who actually speak the languages as well as shifts in the languages themselves. An investigation into ChatGPT’s outputs in several low-resource languages demonstrates the depth of problem.

But language disparities are just one of several concerns as automated moderation becomes more widespread. From the systemic suppression of content from Palestine to the repeated misclassification of LGBTQ+ content as adult or explicit material, these varied examples demonstrate the risks of overreliance on automated moderation—and the need for stronger safeguards.

Fortunately, there is a great deal companies, policymakers, and civil society can do to help ensure that highly automated systems operate in ways that respect human rights, minimize predictable harms, and provide meaningful accountability when they fail. If companies are going to continue relying on automation to moderate users’ speech—and there is little reason to believe they won’t—then accountability must evolve alongside these technologies.

That work should start with publicly committing to the Santa Clara Principles 2.0. Specifically:

1. Automated technologies should help, not replace, human moderators. For example, automated systems can help flag and prioritize content for review, while humans can interpret context, handle sensitive cases, and refine system performance. This investment is particularly important during the weeks immediately before and after an election.
2. Companies must be transparent about when and how automation is used in content decisions.
3. Companies must regularly audit their automated systems for bias, with particular attention to low-resource languages, vulnerable and marginalized communities, and conflict zones.
4. Companies should strengthen personal data protection guarantees and policies. Automated content moderation often relies on personal data and profiling, so companies should think carefully about what data these systems use, what’s necessary and proportionate and how their systems shape moderation outcomes.
5. Users must have the ability to appeal, and to provide context when they believe human or automated moderation decisions have wrongfully removed their content. Appeals should be promptly evaluated and decided by human moderators.
6. Companies should regularly assess the human rights impact of their moderation decisions, and issue public statements of the results
7. If they rely on third-party vendors, companies should carefully (and regularly) audit those vendors for compliance with these same principles
8. Lawmakers should avoid promoting and passing legislation that effectively or explicitly mandates automated moderation systems
9. Policymakers should also refrain from attempting to dictate platforms technical and design choices to favor or disfavor particular expression.
10. Policymakers should work to educate the public about how their personal data may be collected and used in electoral contexts.

These recommendations understand that automated content moderation isn’t just a technical problem for clever engineers and product teams to solve. Because content moderation shapes public discourse and fundamental rights, its design and oversight must respond to the concerns of policymakers, civil society, independent researchers, and the communities most affected by these systems.

One other distinct and enduring problem makes transparency, checks, and accountability all the more urgent: too few companies have far too much power over our political communication and organizing. Social media gave people the ability to wrest some control over political messaging away from media oligopolies. But we’ve met the new boss and while it’s not the same as the old boss, the song is pretty familiar. The real promise of the web to democratize communication is not likely to be met by social media oligarchs, no matter how sophisticated their content moderation practice.