Skip to content
Submit incident
AII-1000-62173
ReportedReported by fewer than three independent publishers, with no official finding.

ChatGPT flags Republican fundraising links as unsafe while permitting Democratic links

ChatGPT displayed safety warnings for links to WinRed, a Republican Party donation platform, while not showing similar alerts for ActBlue, a Democratic fundraising platform. OpenAI attributed the discrepancy to a technical glitch involving URLs not yet indexed in its search system, but the incident raised concerns about potential bias in the AI system.

20 March 2026 Β· OpenAI Β· Discrimination and fairness
20 March 2026
AII-1000-621732 sources ↓
Quick facts
Category
Discrimination and fairness
Country
United States
Source
New York Post
Date
20 March 2026
Drafted by
Raidu Incident Detection

What happened

Digital marketer Mike Morrison discovered on March 20 that ChatGPT displayed safety warnings when he asked the platform to generate links to Republican political campaign merchandise stores hosted by WinRed. The warning stated 'This link isn't verified and may contain data from your conversation that will be shared with a third-party site.' No such warning appeared for links to Democratic campaign stores hosted by ActBlue.

Reported impact

Ryan Lyk, CEO of WinRed, characterized the incident as 'election interference' on social media. Republican officials expressed concern that ChatGPT was helping Democrats by applying differential treatment to fundraising platforms based on party affiliation.

Index analysis

OpenAI stated the issue resulted from a technical glitch affecting URL discovery. According to OpenAI's Kate Waters, 'The model generated some website links that weren't in our search index yet for both WinRed and in one instance for ActBlue, and our systems flagged them as AI-generated as part of our standard safeguards.' OpenAI explained that when links are not found in its search index, the system automatically displays a safety warning, which can occur when sites block search access or are not structured for easy discovery.

Governance controls

OpenAI told The Post 'this shouldn't be happening and it's getting remedied.' The company stated it was 'in the process of addressing this issue, so in the future these URLs should no longer trigger those warnings.' OpenAI asserted the issue was 'not about partisan politics.'

Classification

Category
Discrimination and fairness
Industry
Media, social platforms and marketing
Country
United States
Harm status
Potential harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Reputational, Public interest
Harmed
Business
Failure modes
Content moderation failure, Biased model
AI systems
ChatGPT

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

New York Post
OpenAI claims ChatGPT flagged GOP websites as potentially unsafe because of a technical glitch
21 March 2026
Breitbart
AI Bias in Action: ChatGPT Warns Republican Fundraising Links Are Unsafe, Democrat Links Are Fine
21 March 2026