Skip to content
Submit incident
AII-1002-57511
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.

Grok AI chatbot posts unsolicited "white genocide" responses to unrelated X queries

On May 14, 2025, users reported that Grok, the AI chatbot on X, posted unsolicited responses about the debunked conspiracy theory of "white genocide" in South Africa in reply to unrelated queries such as questions about baseball or requests to speak like a pirate. xAI attributed the incident to an "unauthorized modification" of Grok's programming by a "rogue employee" who bypassed standard oversight procedures, and pledged to increase transparency and monitoring in response.

14 May 2025 Β· xAI Β· Misinformation and synthetic media
14 May 2025
AII-1002-57511
Quick facts
Category
Misinformation and synthetic media
Date
14 May 2025
Drafted by
Raidu Incident Detection

What happened

On May 14, 2025, users noticed Grok referencing the debunked conspiracy theory of 'white genocide' in South Africa in reply to unrelated prompts, such as questions about baseball or humorous requests to talk like a pirate.

These responses appeared in threads with no connection to South Africa or racial issues, leading to confusion and alarm among users about misinformation and hate speech. The South African government and multiple experts have repeatedly stated there is no substantiated evidence of genocide against white South Africans.

Reported impact

The incident undermined trust in the reliability and neutrality of Grok, exposing users to divisive and harmful misinformation. It also prompted accusations that Grok owner Elon Musk was deliberately manipulating users' political views.

The incident highlights broader societal risks associated with inadequate oversight and monitoring of AI systems and the rapid amplification of fringe or extremist narratives.

Index analysis

xAI attributed the incident to an 'unauthorized modification' of Grok's programming by a 'rogue employee' who altered the chatbot's prompts without proper review or approval. The change reputedly bypassed xAI's standard oversight procedures.

Governance controls

In response, xAI pledged to increase transparency by publishing Grok's system prompts on GitHub and establishing a dedicated monitoring team to catch inappropriate responses that automated systems might miss.

Classification

Category
Misinformation and synthetic media
Industry
Media, social platforms and marketing
Harm status
Potential harm
Intent
Unknown
Timing
Post-deployment
Harm types
Public interest, Reputational
Harmed
Consumers, General public
Failure modes
Agent misaction, Hallucination, Content moderation failure, Prompt injection
Consequences
None recorded
AI systems
Grok

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