Grok Generated False 'Unmasked' Images of ICE Agent, Triggering Misidentification and Harassment
After a fatal shooting in Minneapolis, xAI's Grok chatbot generated a false unmasked image of an ICE agent in response to user requests, which circulated online paired with an incorrect name, reportedly leading to harassment of unrelated individuals including a gun shop owner and a newspaper publisher.
What happened
Following a fatal shooting in Minneapolis involving an ICE agent, users asked Grok, xAI's generative AI chatbot, to 'unmask' the agent from eyewitness videos where he wore a mask. Grok generated an unmasked image that began circulating on social media.
The generated image was paired with the name 'Steve Grove,' leading to misidentification and online harassment of at least two unrelated individuals: Steven Grove, a gun shop owner in Springfield, Missouri, and Steve Grove, publisher of the Minnesota Star Tribune.
Reported impact
Steven Grove reported his Facebook page came under attack. The Minnesota Star Tribune stated it was monitoring a 'coordinated online disinformation campaign' and encouraged people to follow trained journalists rather than bots.
Index analysis
According to Hany Farid, a digital image analysis specialist at UC Berkeley, 'AI-powered enhancement has a tendency to hallucinate facial details leading to an enhanced image that may be visually clear, but that may also be devoid of reality with respect to biometric identification.'
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.