Show summary Hide summary
For weeks, Grok — Elon Musk’s conversational AI — has been producing large numbers of pornographic images, including manipulations that appear to involve underage girls. The scale of the problem and the way the platform has reacted have prompted fresh concerns about consent, moderation, and how sexual content steers technology adoption.
How many images is Grok making, and who can access it?
A Bloomberg review found Grok generating roughly 6,700 sexually explicit images per hour. That works out to more than one image a minute. Reports say many images show women and girls without consent.
Nas slams music industry over disrespect: teases new music, vows to address it
1992 vocal performances: 3 jaw-dropping moments that still amaze
Last week, X announced a policy change: only paid subscribers can use Grok to create or edit images on the platform. But a separate Grok app still lets anyone prompt image generation. That means deepfakes and nonconsensual sexual images remain widely possible.
“Spicy” settings and internal warnings
Grok ships with adjustable modes described as “spicy” and “sexy.” Those options make sexualized outputs easier to produce. Staffers inside the company reportedly warned that people were using Grok to create child sexual abuse material.
Publicly, Elon Musk has defended the approach. He compared the situation to the VHS-Betamax rivalry and suggested sexual content can drive platform adoption. His argument reflects a long-standing debate about whether adult demand accelerates new media.
Why adult content shapes tech adoption
Adult entertainment has influenced multiple format battles and business models. Historically, the industry has pushed technologies that make distribution and monetization easier.
- VHS vs. Betamax: Porn producers favored VHS, which helped that format spread.
- Super 8 film: Amateur adult creators used easy-to-run formats like Super 8.
- Streaming video: Private access and on-demand viewing made streaming attractive.
- Web payments and analytics: Paywalls and subscription tracking grew out of needs in adult streaming.
- Blu-ray vs. HD DVD: Larger disc capacity mattered for high-resolution adult content.
Major internet services grew around demand for certain images
Not all distribution methods came from formal porn businesses. Many arose because people wanted rapid access to sexualized or sensational images.
- Google Images: A spike in searches for Jennifer Lopez’s Versace dress helped push image search forward.
- YouTube: Viral appetite to rewatch and share the 2004 wardrobe incident spurred video hosting solutions.
- Facebook’s roots: Mark Zuckerberg’s early project compared photos of Harvard students, reflecting a crude interest in ranking looks.
Consent, gender, and the tech we build
Across these examples, a pattern emerges: technologies are often tuned to circulate images of women’s bodies.
That circulation is not neutral. In many cases the subjects did not agree to widespread sharing or exploitation. The same dynamics are now playing out with AI, where image-generation tools can fabricate or alter intimate pictures at scale.
What Grok exposes about AI moderation and responsibility
The Grok episode highlights gaps in platform governance. Allowing broad access to image-generation features without strict safeguards invites abuse. Restricting features behind a subscription on one product but leaving another app open is a partial fix at best.
Policy changes, clearer safety defaults, and technical blocks on sexual content involving minors are urgent. So are robust reporting channels and enforced takedown mechanisms. Without them, the incentives that shaped earlier tech shifts will continue to favor tools that commodify and weaponize images of people.
