The Smoke Screen: Why we’re talking about Grok bikinis while X plunders our data
While mainstream media is fixated on the news that the AI "Grok" is now restricted from generating bikini photos (a reaction to pressure from the EU and UK's Ofcom), a much larger drama is unfolding in the shadows.
The Real Scandal: Your data now belongs to X
As of today (Jan 15, 2026), the new Terms of Service (ToS) are officially in effect. This means: Everything you share is being ruthlessly harvested for AI training. Meanwhile, the platform is drowning in a sea of scams, bots, and deepfakes.
The Deep Dive: What the new X ToS actually mean for you While everyone stares at the "bikini ban," X has quietly weaponized the fine print. It’s essentially a digital expropriation of your privacy:
Forced Training for Grok: You grant X a worldwide, irrevocable right to use every post—text, photo, or video—to train their AI models. You are no longer a user; you are free fuel for Musk’s ambitions.
The "Sub-License" Trap: X can pass your content to third parties. Your private photos could end up in databases you didn't even know existed.
No Real "Opt-Out": The legal loopholes are so wide that simple settings offer almost no real protection anymore.
Algorithmic Engineering: Speak the truth, get throttled
X’s talk of "Free Speech" is pure cynicism. I’ve experienced it firsthand: I posted excerpts from my anti-fake tool setup to expose scams and warn users. The result? Immediate throttling.
Since the switch to the fully automated Grok recommendation system in late 2025, critical content isn't necessarily deleted—it’s made invisible. If you threaten the bot-infested house of cards, your reach drops to zero. This isn’t freedom; it’s algorithmic censorship.
From Free Speech Haven to a Sanctuary for Harassment
It is bitterly ironic. Under the guise of free speech, floodgates were opened that allow non-consensual sexualized content and deepfakes of private individuals and minors. Global reactions (Malaysia's block, UK investigations) show: the line has been crossed. The current bikini filters feel like a band-aid on a gunshot wound.
My Personal Final Line
I’ve seen enough. On January 12, 2026, I deleted the app. Any platform that treats its users like data-slaves for AI shouldn't be surprised by the mass exodus.
X has been removed from my useful links page. I’m moving on. I hope to see you soon on Bluesky or my independent link hub: [
The "Bikini Update": A Poor Excuse for Safety
Coinciding with the new ToS taking effect on January 15, 2026, X hastily tried to calm the waters. Grok is now supposedly blocked from generating bikini images of real people. But a look behind the curtain reveals a total failure:
Leaky Filters: With simple prompting tricks (e.g., using "crop-tops" instead of "bikinis"), these blocks are already being bypassed on day one.
Musk’s Loophole: While real people are (theoretically) protected, generating explicit images of "imaginary persons" remains explicitly allowed.
Pure Outside Pressure: These changes aren't a change of heart; they are a panic reaction to investigations by the EU, Canada, and the UK’s Ofcom.
X is pretending to make the platform safer. In reality, they are just saving their own legal skin while the deepfake machinery at the core keeps running.
Midnight Update: A Message to the Bots in Morristown
It’s nearly 10 PM. While X has officially implemented its new AI-harvesting Terms of Service today, my analytics are showing a strange light.
Hello, Morristown. While my friend René is being throttled to a mere 3 views on X for sharing the truth, the platform's automated crawlers from New Jersey and Virginia are working overtime to scan every word on this blog.
To the humans reading this: You are witnessing the "Dead Internet Theory" in real-time. The platform suppresses human reach while its own bots swarm the web to feed the AI machine. To the bots from Morristown: You can scrape my syntax, but you can’t own my intent. I deleted @geliebte_macke because I refuse to be part of an AI training set without consent.
My light is still burning, but it’s no longer fueling your algorithm.
January 16, 2026, 8:30 a.m. CET. Good morning to Chino Valley. I wish you all a wonderful day. Things are going to get exciting for me, but I won't reveal any more than that.
January 16, 2026, 8:38 a.m. CET. Good morning to Boeme and Mount Pleasant as well. I wish you a pleasant day too. However, I must now proceed to work. I am occupied.
In the meantime, feel free to read my other articles on my blog. I've been writing about the problems with fakes on X since April 2025. On June 12, 2026, I left X with a clear conscience. I'm staying in my world. Building my tools in Python, adding new components to my scalable anti-fake tools – without X.
January 16, 2026, 09:10 a.m. CET. Good morning to Tampa. 
Comments