Sources: Meta told political campaigns in March that the company reviews AI-generated images via its fact-checking program, which exempts candidates and parties
Isaac Stanley-Becker / Washington Post :
Context & Ripple Effects
Meta's answer to a question about who reviews AI-generated political images exposes a gap at the center of its moderation design: per sources, the company told campaigns in March that AI-generated images go through its fact-checking program — a program whose policy exempts candidates and parties from rating. The people most able to produce persuasive synthetic imagery are therefore the ones its review pipeline does not cover.
That exemption sits awkwardly next to the rest of Meta's 2023–2024 election playbook: requiring advertisers to disclose AI-altered media in political ads, blocking political advertisers from using its generative ad tools after lawmaker warnings ([[a:845995]]), and building cross-industry labeling standards plus an EU operations center before the June 2024 European elections. Meta later reported AI content was under 1% of fact-checked election misinformation — a figure measured only over content fact-checkers actually see.
First-order effects
- Candidates and parties can post AI-generated images on Facebook and Instagram knowing no third-party fact-checker will rate them, while ordinary users' synthetic images remain subject to review.
- Meta's own disclosure rules split its ad ecosystem: advertisers altering media must label it starting in early 2024, but the exemption means campaign-run pages face no equivalent review of AI imagery.
Second-order effects
- Lawmakers who pushed Meta to bar political advertisers from its generative ad tools now have a concrete counterexample to point to, raising pressure for a disclosure or labeling rule that covers candidates rather than only advertisers.
- Because fact-checking cannot reach politician-posted content, Meta is pushed toward self-executing mechanisms — its proposed AI-content labels on Facebook, Instagram, and Threads become the load-bearing control for exactly the actors the exemption protects.
Third-order effects
- If the pattern holds across platforms, enforcement of synthetic-media integrity in politics migrates from third-party fact-checkers to platform-owned provenance and labeling infrastructure — a shift that concentrates judgment calls about political AI imagery inside the companies themselves.
- The exemption-plus-labeling combination creates an asymmetry other democracies' regulators, watching Meta's EU election preparations closely, may treat as a template problem: identical content faces different scrutiny depending solely on whether its poster is a candidate.
The trend: Political AI-content moderation is moving from third-party fact-checking toward platform-built labeling and provenance systems, because exemptions for candidates leave fact-checkers structurally unable to police the highest-volume synthetic-media risk.