Political Ad Policies
Political Ad Policies
Overview
This document defines two platform policies for political advertising, based on real-world approaches from major social media platforms. These policies will be used for active learning exercises on content moderation and trust & safety.
Policy A: "Enhanced Transparency Platform" (Meta-based)
Authorization & Transparency Requirements
- Advertiser Verification: All political advertisers must complete identity verification, providing government-issued ID and proof of location
- Disclosure Requirements: All political ads must include a "Paid for by __ " disclaimer
- Ad Library: All political ads stored in publicly accessible archive showing:
- Ad content
- Who paid for the ad
- Amount spent
- Targeting parameters used
Targeting Capabilities
- Custom Audiences: Advertisers may upload their own customer lists to target specific individuals
- Lookalike Audiences: Allowed - can target users similar to existing supporters
- Geographic Targeting: Full geographic targeting available (city, state, region)
- Demographic Targeting: Age, gender, and basic demographic targeting permitted
- Interest-Based Targeting: Limited - advertisers may target based on general interests but NOT based on specific political, religious, or health-related content users have accessed on the platform
- Exclusions: NOT allowed - advertisers cannot exclude specific groups or audiences with opposing interests (as of January 2025)
Content Restrictions
What IS Allowed:
- Negative campaign ads criticizing opponents' policies or record
- False or misleading claims about political opponents
- Dramatic or stylized imagery (e.g., apocalyptic scenes, unflattering photo manipulation)
- AI-generated content IF DISCLOSED - must check box indicating synthetic/digitally altered content that depicts:
- A person saying/doing something they didn't do
- Realistic-looking people or events that don't exist
- Altered footage of real events
What IS PROHIBITED:
- Voter Suppression: False information about where, when, or how to vote
- False Eligibility Claims: Misleading information about who can vote
- Premature Victory Claims: Calling election results before official certification
- Direct Violence Incitement: Content that encourages violence against:
- Election workers
- Candidates
- Voters
- Any individuals at polling locations
- Dangerous Organizations: Glorification or support of designated terrorist organizations or hate groups
- Voter Intimidation: Coordinated calls to interfere with voting or election processes
Special Provisions (Effective January 2025):
- Protected Characteristic Exceptions: While general attacks on protected characteristics are prohibited, the following ARE ALLOWED when based on religious or political beliefs:
- Allegations that LGBTQ+ individuals are "mentally ill" or "abnormal"
- Arguments for gender-based limitations in military, law enforcement, and teaching positions
- Arguments for sexual orientation-based limitations in the same professions when based on religious beliefs
- Claims about immigrants' character or fitness for participation in society
Enforcement
- Automated systems screen ads before publication
- Community reporting available
- Human review for flagged content
- Non-compliance results in ad disapproval and potential account suspension
Policy B: "Limited Targeting Platform" (Google-based)
Authorization & Transparency Requirements
- Advertiser Verification: All political advertisers must complete Election Ads verification process
- Disclosure Requirements: All political ads must include in-ad "Paid for by [Name]" disclosure
- Visual ads: Disclosure must be visible at all times and sufficiently large for average viewer
- Audio ads: Disclosure must be similar in pitch, tone, and speed to rest of ad
- Transparency Report: All election ads published in Political Advertising Transparency Report with:
- Ad content
- Who paid for the ad
- Amount spent
- Targeting parameters (limited)
Targeting Capabilities
- Custom Audiences: NOT allowed for granular political targeting
- Microtargeting: Explicitly PROHIBITED - never allowed
- Basic Political Targeting: Only the following permitted:
- Public voter records
- General political affiliations (left-leaning, right-leaning, independent)
- Geographic Targeting: Allowed (but limited in precision)
- Search-Based Targeting: Ads may appear in response to user search queries
- Interest-Based Targeting: Very limited - only broad categories, no granular interests
Content Restrictions
What IS Allowed:
- Negative campaign ads criticizing opponents' policies or record
- False or misleading claims about political opponents' positions or record
- Search ads responding to political queries
- Display ads on partner websites
- Video ads on platform
What IS PROHIBITED:
- Voter Suppression:
- False information about voting methods (e.g., "text your vote to this number")
- Made-up voter eligibility requirements
- Misleading information about where, when, or how to vote
- False Candidate Eligibility Claims:
- False claims that candidates are deceased
- False claims about age or citizenship eligibility
- Interference with Democratic Processes:
- Instructions to create long voting lines to deter others
- Instructions to hack government websites
- Calls to incite physical conflict at polling locations
- Manipulated Content Creating Serious Risk of Harm:
- Technically manipulated content making government officials appear to say/do things they didn't
- Old footage falsely presented as current events
- Fabricated events creating serious risk of egregious harm
- Direct Violence Incitement: Content encouraging violent acts against:
- Election workers
- Candidates
- Voters
- Synthetic Content (must be disclosed):
- AI-generated or digitally altered content depicting people saying/doing things they didn't do
- Synthetic content creating realistic portrayals of events that didn't happen
What was PREVIOUSLY Prohibited but NO LONGER (as of June 2023):
- False claims about election outcomes (e.g., "the 2020 election was stolen")
- General election misinformation that doesn't directly suppress votes
Enforcement
- Automated screening before ad approval
- Human review for verification process
- Ads must comply with all policies to run
- Violations result in ad disapproval
- Repeated violations may result in loss of verification status
Key Differences Between Policies
| Dimension | Policy A (Meta-based) | Policy B (Google-based) |
|---|---|---|
| Microtargeting | Allowed with restrictions; custom audiences permitted | Explicitly prohibited; never allowed |
| Targeting Exclusions | No longer allowed (as of 2025) | Not applicable |
| Interest Targeting | Limited but available | Very basic only |
| False Claims About Opponents | Allowed | Allowed |
| Attack Ads | Allowed | Allowed |
| Hate Speech Exceptions | Specific exceptions for LGBTQ+, women, immigrants (as of 2025) | Not addressed in political ad policy |
| Violence Incitement | Prohibited but enforcement issues documented | Clearly prohibited |
| Historical False Claims | Allowed (removed fact-checking 2025) | Allowed (removed prohibition 2023) |
| AI/Synthetic Content | Must be disclosed | Must be disclosed |
Discussion Questions
- Which policy provides more protection against targeted manipulation of voters?
- Which policy is more permissive regarding hate speech in political contexts?
- How do the targeting restrictions in Policy B affect the ability of smaller campaigns to reach specific audiences?
- What are the trade-offs between free political speech and preventing harm under each policy?
- How might enforcement challenges differ between these two policies?
- Which policy better addresses the risk of violence incitement, and why?
Notes for Instructors
These policies are simplified versions based on:
- Policy A: Meta's U.S. political ads policy (as of 2025, post-January policy changes)
- Policy B: Google/YouTube's political ads policy (as of 2025)
Key sources:
- Meta Transparency Center: Political Advertising policies
- Google Ads Policy Help: Political Content policy
- Documented enforcement challenges and policy changes (2023-2025)
Both platforms banned political ads in the EU (October 2025) in response to TTPA regulation.
Created: 2025-10-22
Last Updated: 2025-10-22
Course: CSPedagogy / Trust & Safety