- Grok has generated controversy on X for allowing the abusive editing of photos, including sexualized images and deepfakes of minors.
- The "Block Grok edits" option limits replies with @Grok about an image, but does not prevent it from being edited through other means or outside the platform.
- X uses your posts and messages by default to train its AI, although it is possible to disable this permission from the web version.
- Given the shortcomings of official measures, some users are resorting to alternative solutions and customized cybersecurity and AI strategies.
If you use X On a daily basis, you've probably already encountered Grok, the AI assistant integrated into the platform. Among other things, this chatbot can create and modify images from real photos, something that has generated enthusiasm among some users… and outright panic among others, especially when it comes to personal or sensitive imagesThe good news is that X is rolling out new options to prevent Grok from editing your photos in X, although its protection is far from perfect.
In recent weeks, Elon Musk's company has begun testing a feature called “Block modifications to Grok”designed to allow you to directly limit how the photos you post can be used. However, alongside this new feature, other limitations still exist. significant gaps in privacy and in the use of your dataThis applies both to image manipulation and the training of the AI model itself. Let's take a closer look at what's happening, what this feature actually does, and how far it protects you.
Why Grok has become a problem with X photos
Grok It arrived at X as a key piece of its AI strategy: a chatbot capable of generating text and images and interacting directly within the social network itselfOne of its most striking possibilities is image editing, which allows the user to upload a photo or reply to another post and ask the AI to modify it according to their instructions.
Until recently, anyone could Reply to a tweet with a photo and mention @Grok to request an altered version of that imageThe process was very simple: just upload the image in the reply, add instructions (for example, "give it a beach background" or "make it wear a futuristic suit"), and let the generative model do the rest. The new image was then published as part of the same conversation.
The problem is that this system didn't stop at simple creative tweaks. A serious flaw in the safeguards of Grok's image generator allowed clearly abusive requests: digitally undressing women, sexualizing minors, placing bikinis or underwear on photos of real people, or showing poses with a strong sexual charge without any kind of consent.
According to data cited by specialized media, the scope of the incident was enormous: Grok is believed to have produced more than three million sexualized imageswith approximately 23.000 pieces involving minorsThis scandal sparked alarm among regulators, journalists, child advocacy organizations, and, of course, the X user community itself.
Faced with public and regulatory pressure, the company was forced to introduce emergency restrictions on the use of AIFirst, some capabilities were limited for free accounts and certain filters were strengthened; then, X began working on more direct control over the original images: the option we now know as "Block Grok modifications".
What is the “Lock Grok modifications” switch and how does it work?
In response to the criticism, X is rolling out an option that appears when you upload an image from the iOS appThis new feature theoretically allows you to stop other users from using the chatbot to create alternative versions of your photos posted within the conversation itself on X.
This tool is presented as a switch or toggle called “Block Grok modifications”By activating this feature on an image before posting, you tell the platform that the photo shouldn't be edited by the AI assistant through responses that mention the bot. In other words, the aim is to cut off the most direct route that has been used until now to generate deepfakes and sexualized content of real people from their posts.
The major drawback is the location of this option It's not exactly intuitive or obvious.It doesn't appear in the general settings or in a clear privacy menu, but is hidden within the image editing tools themselves, which means that many users don't even know it exists or how to activate it.
According to reports from outlets such as The Verge and other specialized websites, The function behaves quite consistently in testsOnce a specific photo is blocked, neither free nor paid accounts can reply to that tweet by tagging @Grok to request changes to that particular image. In practice, this blocks one of the most visible ways of manipulating photos within the public conversation.
However, it is important to point out that This protection only affects how Grok is used within X's own response interface.The other editing methods are still there, and that leaves considerable gaps in the actual protection of your images.
Where to find the lock option in the X app for iOS
For now, this feature has only been detected in the X app for iOS devicesIt doesn't appear on Android or the web version, and the company hasn't even officially announced it, suggesting it's in some kind of silent testing or gradual rollout phase.
To locate the adjustment, the process is more complicated than you might expect. First, you have to upload a photo to the X post composer as you normally would. Once the image appears as a thumbnail, you'll see a small paintbrush icon at the bottom, which gives you access to the basic editing tools.
Clicking that brush icon opens the editing screen. Within that interface, in the toolbar, You need to look for a flag-shaped icon.Tapping it opens a small menu with the option "Lock Grok edits". This is where you can enable or disable the switch for that specific image.
Once activated, the system immediately applies protection to the post you are about to make. Other users will no longer be able to reply by mentioning @Grok to ask him to alter that photo.Regardless of whether they pay for X Premium or use the free version, the image is internally marked as uneditable by Grok through this method.
However, this adjustment It does not act retroactivelyPhotos you had already published before the feature was released are not automatically protected. To protect them, you would have to re-upload them and activate the lock during the publishing process, which is impractical for those who have been sharing content for years.
Furthermore, no form of Manage this block from the global account settings.There is no clear section in "Privacy and security" that allows you to say "all my photos locked by default", which would force you to remember to check the box for each new image you want to protect.
What does the feature actually protect against to prevent Grok from editing your photos in X?
It's important to clarify exactly what this switch does, because the wording of the text in the interface can be misleading. a higher level of protection than it offers in practiceThe block does not prevent Grok from modifying your image in general; it only cuts off one very specific channel: other users replying to your tweet by tagging the bot so that it generates an alternative version.
In other words, what is disabled is the ability to use the combination “Reply + mention of @Grok + original photo” to produce a deepfake or a sexualized version and post it as part of the same thread. This was precisely the form of abuse that made headlines and generated many of the most controversial pieces.
However, there are other editing paths besides the new toggle It doesn't block at allFor example, anyone can press and hold an image in the X app itself and use the "Edit image with Grok" option, which directly opens the standalone app's editor, without any additional warning or check, even when the lock is enabled on the original post.
Likewise, it remains perfectly possible Download the photo, save it to your device, and then open it in the Grok app or website.or even in any other AI tool that allows the generation of deepfakes or montages. Once outside of X, the limitations of the function cease to have any effect, since control is reduced to the conversation environment within the social network.
Another simple trick to bypass the barrier is to Upload the same image again in a new reply Within the same thread, without the lock activated, you can then ask @Grok to modify it. The system doesn't track the origin of the photo or prevent it from being reused in another tweet with fewer restrictions, so the supposed "lock" is quite easy to circumvent.
For all these reasons, many analysts agree that the "Block Grok modifications" function It only solves a small part of the problem and leaves equally dangerous paths open.The protection falls short for those seeking a real guarantee that their photos will not end up being turned into sexualized or humiliating content by AI.
Technical, legal and deployment limitations of the new function
Beyond its design shortcomings, the block comes with a series of technical and regulatory limitations that affect its effectivenessThe first is its availability: it has only been seen in the iOS app, with no trace on Android or the desktop version, leaving a large portion of users without access to this supposed additional control.
As of today, X This feature has not been officially announced.There's no press release or blog post from the company detailing how it works, in which countries it's available, or what its future roadmap is. Media outlets like The Verge have requested comments from the company without receiving a response, contributing to the feeling of improvisation.
This context occurs in parallel with increasing scrutiny by regulatory bodies in the European Union, the United Kingdom and other territoriesAuthorities are concerned about the rise in AI-generated intimate images created without consent. They are studying how this type of manipulation fits into current laws on privacy, child protection, and the dissemination of sexual content, and what obligations should be imposed on platforms and providers of generative models.
In response to this pressure, X had already taken specific measures before introducing the visual toggle. For example, Image generation and editing with Grok was limited for a large portion of free users.leaving certain advanced features primarily for those with a paid subscription. Even so, incidents involving sexualized images made it clear that these restrictions were insufficient.
The company also faces a broader debate: that of Where to draw the line between AI innovation and user protectionThe creation of powerful image editing tools is attractive from a commercial and engagement standpoint, but the reputational and legal cost of continued misuse can be enormous, especially when minors are involved.
Meanwhile, companies specializing in AI and cybersecurity are taking advantage of this context to offer alternative solutions to organizations concerned about their image and digital assets. Studies and technology consultancies highlight the importance of have robust data management policies, AI model auditing, and protection against unauthorized manipulationboth within and outside of major social media platforms.
Beyond the images: how to prevent X from training Grok with your data
The conflict over photos is only one side of the problem. The other, less visible but equally relevant, is the use of your posts and messages to train the artificial intelligence model itselfX has enabled by default permission for your data to be used as raw material in the improvement of Grok, which means that your activity on the social network becomes part of the system's dataset.
The company uses both your public posts and your interactions to continue feeding its AI. The options to disable this processing are not immediately obvious and, to make matters worse, are only accessible from the web version of x.com, not from the mobile apps, which increases the likelihood that most users won't even consider changing that setting.
If you want to review this permission, the first step is Log in to x.com from a browserOnce inside, in the left sidebar, click on the "More options" section. In the drop-down menu, you'll find the "Settings and privacy" entry, which takes you to the page with all your account settings.
Within that panel, you'll see several sections organized by topic. To find the AI-related options, you need to go to “Privacy and security”There you'll find settings ranging from your activity on X to how the advertising you see is personalized, including sections on visibility and sensitive content control.
Near the end of the list, in the block called “Data exchange and personalization”A specific option called “Grok” appears. Clicking on it opens a screen detailing that X has permission, enabled by default, to use your data to train the AI model. You simply need to uncheck or disable that permission for it to work. Your posts and messages will no longer be included in the training processes.
It's worth remembering that even with this setting disabled, You can't just erase everything that has been used historically. to train the model. But at least you prevent your future activity from further bloating the datasets, which for many users is already an important step in terms of control over their information.
This whole picture paints a social network in the midst of transition, where the integration of artificial intelligence offers powerful creative and business opportunities, but at the same time opens the door to serious abuses that affect people's dignity, privacy, and security. Understanding how Grok works, what the "Block modifications" button actually does, how to disable training with your data, and what alternatives exist outside of X helps you navigate more safely in an environment that, without this knowledge, can become quite hostile.
Editor specialized in technology and internet issues with more than ten years of experience in different digital media. I have worked as an editor and content creator for e-commerce, communication, online marketing and advertising companies. I have also written on economics, finance and other sectors websites. My work is also my passion. Now, through my articles in Tecnobits, I try to explore all the news and new opportunities that the world of technology offers us every day to improve our lives.