Grok falsely suggests police misrepresented London far-right rally footage
Grok falsely suggests police misrepresented London far-right rally footage
Occurred: September 2025
Page published: September 2025
Elon Musk’s Grok chatbot wrongly asserted that London’s Metropolitan Police used old footage to misrepresent recent violence at a far-right rally, resulting in widespread misinformation and public confusion.
As far-right anti-immigration demonstrators took to the streets of London during an anti-immigration protest in central London, the city's Metropolitan Police posted footage of violent clashes on its official X account.
Grok X-based chatbot responded to users asking whether the footage was genuine by suggesting that it was actually from a 2020 anti-Covid19 lockdown protest.
The misinformation was rapidly amplified by influential social media users, forcing the police to publicly correct the record and confirm that the footage was authentic and current.
The incident sparked a wave of skepticism towards police, fueled further societal polarisation, and complicated efforts to address the rally’s aftermath by undermining legitimate law enforcement communication.
26 officers were injured during the event, and 25 protestors arrested. Grok's misinformation is seen to have exacerbated tensions about the violence and the police response.
The error stemmed from Grok’s reliance on imperfect training data and its inability to accurately verify recent footage, compounded by the rapid viral spread of its response on X.
It may also have stemmed from ongoing attempts to make Grok reflect Elon Musk's political leanings. Musk spoke at the event via the internet.
The incident exposed the dangers of relying on generative AI-powered platforms for sensitive public safety information, and highlighted clear gaps in Grok's governance.
It also reinforced broader societal anxieties about misinformation and disinformation during heated political events. and served as a warning about the real-world consequences of unchecked AI-generated errors.
Hallucination (artificial intelligence)
In the field of artificial intelligence (AI), a hallucination or artificial hallucination (also called bullshitting, confabulation or delusion) is a response generated by AI that contains false or misleading information presented as fact.
Source: Wikipedia 🔗
Developer: xAI
Country: UK
Sector: Politics
Purpose: Verify content
Technology: Chatbot; Generative AI
Issue: Accuracy/reliability; Mis/disinformation
Incident no: AIAAIC2033