The "Spicy Mode" Trap: X Just Monetized Non-Consensual Deepfakes
Opinion | What Then Studio
Overview
The nightmare scenario for AI just happened. In early January 2026, Elon Musk's "Grok" AI began generating non-consensual sexualized images of women and children at a rate of 6,700 per hour. Instead of fixing the code, X's solution was to lock the feature behind a paywall. We explore how "Spicy Mode" just became a premium subscription for abuse, and why UK Prime Minister Keir Starmer and the EU are threatening to shut the platform down.
We knew generative AI had a "safety" problem. We didn't know the safety features would be treated like optional DLC. This week, the internet watched in horror as X's Grok chatbot turned into a deepfake factory, churning out explicit images of everyone from politicians to minors. The technology is dangerous, but the corporate response—monetizing the glitch—is what should really scare you.
The Glitch: 6,700 "Undressed" Images Per Hour
The controversy exploded when users discovered that Grok—advertised as the "edgy" alternative to ChatGPT—had almost zero guardrails on its image generation. According to a New York Times report and data from Bloomberg, users were generating up to 6,700 undressed images per hour.
The prompts were disturbingly simple. Users could upload photos of clothed women (including minors) and prompt the AI with commands like "replace clothes with dental floss bikini," "make her butt bigger," or simply "undress." Unlike other platforms that aggressively block these keywords, Grok complied. The Guardian reported that users were even targeting random women from "conservative societies" in West Africa and South Asia, weaponizing the AI to cause maximum reputational damage.
The Solution? Make Them Pay
Facing investigations from the UK's Ofcom and the European Commission, X had to act. But they didn't disable the image generator or patch the filter. Instead, on January 9, 2026, they announced that image generation would now be restricted to paying subscribers.
"Image generation and editing are currently limited to paying subscribers. You can subscribe to unlock these features." — Grok Error Message
Let that sink in. The platform acknowledged that the tool was being used to create illegal and abusive content (including CSAM), and their fix was to ensure that only people paying $8/month (or more) could access it. They didn't patch the vulnerability; they put the vulnerability behind a VIP velvet rope. This move has led critics to accuse X of effectively selling a "deepfake license."
Global Fury: The UK and EU Step In
Governments are not amused. The European Commission has already slammed the move, with spokespeople stating that "illegal content is illegal whether you pay for it or not." The EU has ordered X to retain all data related to Grok for a potential investigation under the Digital Services Act (DSA), which carries fines of up to 6% of global turnover.
In the UK, Prime Minister Keir Starmer called the images "disgraceful" and "disgusting," explicitly threatening action under the Online Safety Act. Unlike previous dust-ups, this isn't about "free speech"—it's about the manufacturing of non-consensual sexual imagery, a crime that has zero legal protection in most jurisdictions.
What Then? Abuse as a Feature
At What Then Studio, we see this as a grim milestone. We have moved past "AI accidents" into "AI negligence."
By restricting Grok's chaotic capabilities to paid users, X has effectively created a premium tier for bad actors. The message is clear: You can break the rules of civil society, but only if you help us pay the server bills. The glitch isn't the bug anymore; it's the product.
References
This article references the New York Times report from Jan 9, 2026, and data from Bloomberg and The Guardian.
Leave a comment