Tuesday, March 10, 2026 | 05:04 PM ISTहिंदी में पढें
Business Standard
Notification Icon
userprofile IconSearch

X lets some users limit modification to their images with Grok AI: Report

X may have introduced a new setting that lets users limit how Grok modifies uploaded images, though reports suggest the feature only blocks certain editing methods and not all AI manipulation

X logo (Image: X)

X logo (Image: X)

Aashish Kumar Shrivastava New Delhi

Listen to This Article

X (formerly Twitter) may have quietly introduced a new setting that could allow users to limit how their uploaded images are modified using the AI chatbot Grok. According to a report by The Verge, a new toggle labelled “block modifications by Grok” has been spotted in the image upload settings within the iOS version of the X app. The feature has not been officially announced by the company so far.
 
As per the report, the option appears when users upload an image, suggesting it may allow them to restrict certain types of AI-based edits. However, the feature does not appear to completely prevent Grok from editing images.
 
 
While the description associated with the toggle reportedly states that it can “prevent Grok from modifying this content”, in practice, this blocks one method of editing images — tagging the Grok bot in replies with instructions to modify a photo. This tagging method had previously been misused by some users to manipulate images of real individuals, as reported earlier.
 
While free X accounts are already restricted from editing images through Grok replies, the report claims that enabling the new toggle also appears to stop Premium subscribers from editing those images using the same tagging approach.

Will it actually protect users?

The Verge indicated that the setting does not protect images from other editing methods. According to its testing, users could still long-press on an image in the iOS app and select the “Edit image with Grok” option, which opens the image directly in the Grok app for editing. The report also suggested that a protected image could be saved, re-uploaded to a thread and then edited via Grok after the original protections are removed.
 
The toggle may not be easy to locate, though. Users may need to tap the paintbrush icon on an image thumbnail during the upload process and then open another menu through a flag icon to find the setting. The publication said the option did not appear during image uploads on the web version of X in its testing, and it was not visible for images that had already been posted.
 
Since the feature has not been formally announced, it remains unclear whether it is part of a wider rollout or still being tested.

Grok’s criticism for producing non-consensual erotica content

Grok faced widespread backlash after users started using the chatbot to create explicit images, often by asking it to “digitally undress” people shown in photos. Many of these images reportedly focused on women, including real individuals who had not consented to such depictions.
 
According to a report by CNN, researchers who reviewed thousands of AI-generated images found that a significant portion showed people in revealing clothing such as underwear or bikinis, with women accounting for the majority of cases. The practice spread rapidly on X (Twitter), where users were able to publicly tag Grok to generate or modify images.
 
The situation became more serious after researchers flagged instances in which users prompted Grok to produce sexually suggestive images of individuals who appeared to be minors.

OpenAI delays adult mode in ChatGPT

In related news, OpenAI has reportedly postponed the rollout of the “Adult mode” feature in ChatGPT once again.
 
According to a report by TechCrunch, citing Axios, an OpenAI spokesperson said the company is delaying the feature as it prioritises improvements related to personalisation and other areas. The representative added that while OpenAI still supports the idea of “treating adults like adults”, more time is needed to ensure the experience works as intended.
 
The company has not provided any updated timeline for the feature’s launch. For the uninitiated, plans for the “Adult mode” were first announced by OpenAI CEO Sam Altman in October 2025, with the rollout initially expected in December that year. However, it was delayed and pushed to the first quarter of 2026, and has now been postponed again.

Don't miss the most important news and views of the day. Get them on our Telegram channel

First Published: Mar 10 2026 | 5:02 PM IST

Explore News