The Risks of AI in Unmasking: A Misidentification Sparks Backlash
An original still image from an eyewitness video shows the masked ICE agent who shot Renee Nicole Good (Left). Users on social media “unmasked” the agent using Grok (Right). Experts warn AI cannot “unmask” individuals. NPR is publishing both images to show how AI is being used to manipulate images of news events.
Screenshots by NPR/Image by Courtney Theophin/NPR
In a recent incident in Minneapolis, the shooting of Renee Good, 37, by an ICE agent ignited a storm of controversy and misinformation. The agent, whose face was obscured in eyewitness footage, was seemingly “unmasked” on social media through the use of xAI’s generative AI chatbot, Grok. This AI-generated image was widely shared, yet experts caution against the reliability of such technology for accurate identification.
Despite the agent’s identity being concealed in the original videos, Grok’s output, which was prompted by users on X, suggested an unmasked likeness. NPR has chosen to publish both images to highlight the potential for AI to distort factual representations of news events, while experts emphasize the unreliability of AI in revealing true identities.
Hany Farid, a professor at the University of California, Berkeley, who focuses on digital image analysis, addressed this issue: “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.”
The AI-generated image, however, led to a case of mistaken identity, resulting in the circulation of the name Steve Grove, which caused unwarranted outrage against two individuals unaffiliated with the event. One individual, Steven Grove from Springfield, Mo., found his social media under attack. He clarified, “I never go by ‘Steve,'” and denounced any connection to the incident, citing his distinct appearance and location.
The second individual affected was Steve Grove, the publisher of the Minnesota Star Tribune, which released a statement to combat what they described as a “coordinated online disinformation campaign.” The Star Tribune urged readers to seek verified news from trained journalists, emphasizing the importance of reliable sources over bot-generated content.
Meanwhile, the Star Tribune, alongside NPR, reported that the ICE agent involved was named Jonathan Ross. Court documents indicate Ross had previously been involved in a separate incident in Bloomington, Minn.
This article was originally written by www.npr.org
