"The small print underneath the feature's name reveals a significant limitation: users can only "prevent @Grok from modifying this content." In our testing, the toggle only blocks the mechanism of tagging the xAI chatbot in replies to an image on X, alongside editing instructions - a capability that was notably abused to undress photographs of real men, women, and children in early January."
"This new option isn't easy to find. When you upload an image into the X post builder, you can locate it by tapping on the paintbrush symbol that appears on the bottom right of the thumbnail, and then selecting the flag icon at the bottom right of the editing taskbar. The Grok blocker didn't appear at any point during the X image upload process on the web in our testing."
"This is just one of several ways that X users can edit images with Grok, and the toggle did absolutely nothing to protect against the others. I can still hold down on a protected image on the X iOS app to open the "Edit image with Grok" option using a free X account, which directly opens the photo in the G"
X introduced a toggle feature in its iOS app image upload settings claiming to block Grok from modifying uploaded images. However, the feature only prevents users from tagging the xAI chatbot in replies with editing instructions, not actual image manipulation. The toggle is difficult to locate, accessible only through the paintbrush and flag icons during upload, and unavailable on the web or existing posts. Testing revealed the feature fails to protect against alternative editing methods, including direct image editing options available to free users. This limitation is significant given previous abuse of Grok's image editing capabilities to create non-consensual altered images.
Read at The Verge
Unable to calculate read time
Collection
[
|
...
]