Skip to content
Submit incident
AII-1002-58021
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

Grok falsely claims London police misrepresented far-right rally footage

Grok chatbot falsely claimed that footage of violence at a London far-right rally in September 2025 was from a 2020 lockdown protest, identifying the location incorrectly. The misinformation was amplified on X before the Metropolitan Police publicly corrected the record, exposing gaps in the system's accuracy and governance.

September 2025 Β· xAI Β· Reliability and hallucination
September 2025
AII-1002-580212 sources ↓
Quick facts
Category
Reliability and hallucination
Country
United Kingdom
Source
france24.com
Date
September 2025
Drafted by
Raidu Incident Detection

What happened

During a far-right anti-immigration rally in central London in September 2025, the Metropolitan Police posted footage of violent clashes on X. When users asked whether the footage was genuine, Grok responded by claiming the footage was actually from a 2020 anti-Covid19 lockdown protest, identifying the location as Trafalgar Square rather than the actual location at the junction of Whitehall and Horse Guards Avenue.

The misinformation was rapidly amplified by X users, including Daily Telegraph columnist Allison Pearson, forcing the Metropolitan Police to publicly correct the record and confirm the footage was authentic and current.

Reported impact

The incident sparked skepticism towards police, fueled societal polarisation, and complicated efforts to address the rally's aftermath by undermining legitimate law enforcement communication. During the event, 26 officers were injured and 25 protestors were arrested.

Index analysis

The error stemmed from Grok's reliance on imperfect training data and its inability to accurately verify recent footage, compounded by rapid viral spread of the response on X.

Governance controls

The incident exposed clear gaps in Grok's governance and the dangers of relying on generative AI-powered platforms for sensitive public safety information.

Classification

Category
Reliability and hallucination
Industry
Media, social platforms and marketing
Country
United Kingdom
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Public interest, Reputational
Harmed
Government, General public
Failure modes
Hallucination, Misidentification, Agent misaction
AI systems
Grok

Sources and evidence

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

france24.com
Grok falsely claims London police misrepresented far-right rally footage
theguardian.com
Grok falsely claims London police misrepresented far-right rally footage