Grok users say it's no longer generating NSFW videos from uploaded images

Update 29/10/25 – 04:14 pm (IST): The frustration is leading some users to leave negative reviews for Grok on the App Store and Play Store. Meanwhile, some users are sharing potential workarounds to beat the moderation hammer.

Original article published on October 29, 2025, follows:

If you’ve been using Grok to create spicy content from uploaded images, you probably noticed something broke over the past day or two. What used to slide through moderation without issue is now getting flagged left and right, and users are pretty pissed about it.

The lightbulb moment hit for one Reddit user just a few hours ago. They’d been reading complaints for days but couldn’t understand what people were so upset about since their own results were still coming through fine. “If this is what people are calling ‘bad’, then how the hell was it before!?” they wondered. Then the moderation hammer finally dropped on them too, and everything that worked yesterday stopped working today.

Turns out xAI didn’t flip a switch for everyone at once. They rolled out these restrictions in waves, which is why some people were still generating content while others were already locked out. One user in the comments laid out how things changed: before October 15th, basically anything went. From October 15 through October 27, you could still get some stuff through if you were lucky, though it became way harder. Now even cartoon breasts get rejected.

Here’s the weird part — the model itself reportedly still acts like it wants to make adult content. Upload a totally innocent photo of two people sitting next to each other, ask for a “normal” video, and Grok decides on its own that they should start making out. So the behavior is still there, but the moderation is blocking it. Makes zero sense.

If user reports are anything to go by, the moderation got even stricter in the last few hours. “All uploaded Images to Video no longer works,” someone confirmed. Some people are even getting weird unprompted results with distorted, monster-like figures showing up.

The timing here is pretty interesting. xAI staff were reportedly working on adult content under something called “Project Rabbit,” while ChatGPT just announced they’re adding erotic features with age verification. Grok literally marketed its “Spicy” mode for explicit content as a selling point, so doing a complete 180 without warning feels like a bait and switch.

Some users think this is more about app store politics than actual legal concerns. “If it was truly about legal and safety issues, they would have NEVER let us have what we did a few weeks ago in the first place,” one person argued. The theory? This is a business move. Get people hooked with loose restrictions, pull back to keep Apple and Google happy, then bring some of it back later with a premium tier once the dust settles.

That said, until xAI explains what’s going on, everything is pure speculation and should be taken with a grain of salt. After all, this could also be a simple moderation glitch that might get resolved shortly. 

We’ll keep an eye out for any further developments and will update this article if and when there’s something to share.