Elon Musk’s Grok Is a Risky Experiment in AI Content Moderation

The unfettered image generation tool on X is a test case for safe AI

Provide news feedback or report an error

Confidential tip?

Send a tip to our reporters

Site feedback:

Take our Survey

By Davey Alba

August 26, 2024 at 4:00 PM UTC

This article is for subscribers only.

A deluge of bizarre computer-generated images swept Elon Musk’s social platform X last week — including violent, offensive and sexually suggestive content. In one, Trump piloted a helicopter as the World Trade Center buildings burned in the background. In others, Kamala Harris wore a bikini, and Donald Duck used heroin. Amidst the online furor, Musk posted, “Grok is the most fun AI in the world!”

By Friday, the shocking images had lost some of their novelty. The volume of posts about Grok peaked at 166,000 posts on Aug. 15, two days after the image generation features were announced, according to the data firm PeakMetrics.