Skip to content
Submit incident
AII-1001-82697
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.Supporting sources are pending verification.

LA Times AI Tool Generates Defense of KKK Within 24 Hours of Launch

The Los Angeles Times' newly deployed AI tool generated a defense of the Ku Klux Klan within 24 hours of launch in response to an opinion column criticizing the organization. The problematic content was removed after being discovered and reported.

4 March 2025 Β· Particle Β· Harmful and abusive content
4 March 2025
AII-1001-82697
Quick facts
Category
Harmful and abusive content
Country
United States
Date
4 March 2025
Drafted by
Raidu Incident Detection

What happened

The Los Angeles Times launched an AI feature called 'Insights,' created by the company Particle, on March 3, 2025. The tool was designed to provide counterarguments and alternative perspectives to opinion columns.

Less than 24 hours after deployment, the AI generated a defense of the Ku Klux Klan in response to a column by Gustavo Arellano about the 100th anniversary of Anaheim removing four KKK members from its city council. The AI stated: 'Local historical accounts occasionally frame the 1920s Klan as a product of "white Protestant culture" responding to societal changes rather than an explicitly hate-driven movement, minimizing its ideological threat.'

The problematic AI-generated content was removed after being discovered and reported by New York Times tech reporter Ryan Mac.

Reported impact

The incident prompted criticism from the L.A. Times Guild, with vice chair Matt Hamilton stating the day before: 'This tool risks further eroding confidence in the news.'

Index analysis

The AI system failed to recognize and refuse to generate content that defended a hate group, instead presenting KKK ideology as a reasonable historical perspective worthy of counterbalancing an op-ed criticizing the organization.

Governance controls

The L.A. Times Guild had warned of such risks prior to the feature's launch, predicting the tool would generate problematic content.

Classification

Category
Harmful and abusive content
Industry
Media, social platforms and marketing
Country
United States
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Public interest, Reputational
Harmed
Business, General public
Failure modes
Agent misaction, Biased model, Content moderation failure, Unsafe autonomous behaviour
Consequences
Content removal
AI systems
Insights

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident