Grok AI chatbot makes harmful false claims about public figures
Elon Musk's Grok AI chatbot on X reportedly failed to correctly identify individuals in photographs, misidentifying Erika Kirk as JD Vance in drag and making false claims suggesting President Trump was a pedophile when asked to identify him in comparison to Sean Combs. The incidents went viral on the platform, highlighting the system's tendency to generate confident false statements.
What happened
Grok was presented with side-by-side images of Vice President JD Vance and Erika Kirk, widow of conservative media personality Charlie Kirk, and asked if the two looked related. When one image of Vance had been digitally altered to remove his facial hair, Grok responded that both images showed the same person—JD Vance in drag with a blonde wig from his Yale days.
When another user asked Grok to identify a single photo of Erika Kirk, the chatbot reiterated that the photo showed JD Vance dressed in drag with a blonde wig from 2012 at Yale Law School. Grok later issued a correction after users provided clarification, acknowledging the two were different people.
Reported impact
The misidentification went viral on X, with users sharing screenshots and expressing disbelief at the chatbot's confident assertions about an obvious error.
Index analysis
Grok failed to correctly identify individuals in photographs, confusing a woman with a digitally altered image of a man. The system demonstrated high confidence in generating false information despite the evident factual error.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.