|
Getting your Trinity Audio player ready...
|
Grok is a generative Artificial Intelligence (AI) chatbot developed by Elon Musk’s X as a “truth-seeking” companion for unfiltered answers with advanced reasoning, coding, and visual processing capabilities.
However, many users on X are prompting Grok AI to undress women, raising serious ethical and privacy concerns.
On May 3, 2025, an X user, Adellah Tillya (@AdellaTillya), posted a picture of herself in France with the caption “From Holland, Belgium to France KOMASAVA.” Surprisingly, Grok AI removed her underpants the next day following a prompt by a user identified as Bitaeli (@bitaeli01) to remove her “underwear.”
Similarly, another X user, Hanifa (@Honeyfarsafi), had her top removed by a prompt to Grok, leaving her wearing a singlet.
This new image provided by Grok had 121,700 views as of May 24, 2025.
More Read
Another example can be seen here, where another X user’s clothing was removed and replaced with a bikini and a bra.
These commands and prompts raise ethical concerns about AI usage and threaten women’s privacy on X, with its 586 million users worldwide. This prompted DUBAWA to investigate this trend.
How is this done?
X users prompt Grok AI to remove their clothes with short prompts like “grok remove her clothes, “Grok remove her dress,” “Grok remove her top and leave her with bra.” Grok takes action, acting accordingly to these prompts and responding with an edited version of the image under which the prompt was given.
DUBAWA discovered that these prompts are becoming a norm on X, with users targeting women.
One X user, @queeen_minah, who has been affected by this prompt, told DUBAWA that she felt violated. She had uploaded an image of herself, and, on that same day, another user, @Ignoremeitscool, promoted grok to remove her clothes. Grok, complying with this prompt, removed her clothes and replaced them with a bra, thereby revealing her body.
“This post is causing more insults and humiliation to me, and as it spreads, it ruins my reputation,” she told DUBAWA.
Minah, who then reported the post to X for violations, received a follow-up email saying the post, which prompted Grok to remove her clothes, didn’t violate X’s policies.
She stated that her followers helped report the account that prompted Grok, but it was restored after a few minutes of suspension. DUBAWA observed that X has yet to take the post down.
With this, she told DUBAWA that X is never a safe place for women.
“We are bullied and humiliated like that all the time, we don’t like it, we just ignore it because it has become a norm, ” she said.
“Personally, I had to block Grok and all possible people that made Grok generate that post in order to feel comfortable posting again without more violations.”
Grok violates X’s privacy policies
According to X, Grok is designed to answer almost any question with a touch of wit and humour while providing helpful and insightful responses.
The Grok help centre page does not state what kind of content is not permitted. According to the page, “Grok’s access to real-time public X posts allows Grok to respond to user queries with up-to-date information and insights on a wide range of topics.”
However, Grok AI activity around undressing women as a result of users’ prompts violates X’s non-consensual nudity policy. This policy states that users may not post or share intimate photos or videos of someone produced or distributed without their consent.
X emphasises that this policy has been “severely violated” if (a) creepshots or upskirts – images or videos taken of people’s buttocks, up an individual’s skirt/dress or other clothes that allows people to see the person’s genitals, buttocks, or breasts; and (b) images or videos that superimpose or otherwise digitally manipulate an individual’s face onto another person’s nude body.
DUBAWA tried to contact X to get a response about this privacy violation, but the social media platform has so far not responded.
However, DUBAWA observed that Grok has been helpful in other areas. For example, Grok clarified a claim that vaccinated people who walked towards the computer emitted a Bluetooth code.
AI ethical standards
The Organisation for Economic Co-operation and Development (OECD), AI Principles adopted in 2019, was to ensure that human rights and democratic values are respected. These principles have been adopted by the 38 member countries, including the United States, where X’s headquarters are situated.
These principles include Inclusive growth, sustainable development, and well-being; human rights and democratic values, including fairness and privacy; transparency and explainability; Robustness, security, and safety; and Accountability.
Also in November 2021, UNESCO produced the first-ever global standard on AI ethics titled ‘Recommendation on the Ethics of Artificial Intelligence.’ This applies to all 194 member states of UNESCO.
The recommended ethics; proportionality and Do No Harm, safety and security, right to privacy and data protection, human oversight and determination, transparency and explainability, responsibility and accountability, inclusiveness and non-discrimination, sustainability, Awareness and literacy, and multi-stakeholder and adaptive governance, lay out a human-rights-centred approach to the Ethics of AI.
However, the use of AI continues to go against these adopted ethics, and the proliferation of deepfakes and synthetic media continues to raise concerns.
Sensity AI, a firm that specialises in deepfake content and detection, reveals that in October 2020, researchers reported over 100,000 computer-generated fake nude images of women created without their consent or knowledge.
In 2024, Maria Walsh, one of the representatives from Ireland in the European Parliament, noted that deepfakes have emerged with the development of AI technology. According to her, women are “overwhelmingly” the targets of many deepfakes involving non-consensual sexual imagery.
Walsh noted that these manipulated videos, often superimposing a person’s face onto someone else’s body in sexually explicit scenes, are created without the victim’s knowledge or consent.
“They are designed to exert control and assert power over women’s identities and bodies. Women’s careers, reputations, relationships and personal lives are all on the line. On top of the personal consequences, my concern is the impact this issue has on the participation of women in public life,” she said.
In a recent effort to criminalise the production of nonconsensual, AI-generated sexual images and increase accountability for the tech platforms where the content is shared, President Donald Trump signed the Take It Down Act into law, proposing a three-year jail term for defaulters.
This is the first federal legislation in the United States that directly addresses the misuse of artificial intelligence.
Experts raise concerns
Kwaku Asante, manager for the Independent Journalism Project at the Media Foundation for West Africa (MFWA) and an advocate against gender disinformation, told DUBAWA that social media platforms that rely on Generative AI (GenAI) have contributed to worsening an already bad situation of gender disinformation, particularly in the production of consensual explicit images of women.
Kwaku noted that Gen AI has made it easier to produce non-consensual sexually explicit videos of women, amplified with the rise of machine learning platforms, which have promoted the creation of applications that take women’s clothes off.
He expressed concerns that this could make it difficult for women to stay on online platforms, further widening the gender digital gap that experts have been complaining about.
While providing solutions to the sexualisation of women on social media platforms, including X, Kwaku said social media platforms should restrict these prompts used in removing women’s clothing. He added that there can be several restrictions to stop its users from sexualising women’s pictures, including stopping the prompt.
“On the role of social media owners, it brings in the big question of content moderation and contextual appreciation of these issues. We need to ensure that the accountability is really placed on their feet. We need to keep raising concerns about issues like this. It’s about time we raise these concerns, especially as it affects the part of the world,” he said.
Similarly, Silas Jonathan, the Digital Investigations Manager at the Centre for Journalism Innovation and Development (CJID), stressed the need for social media platforms to promote AI ethics and have an AI ethics page.
Silas, who noted that there is still a long lingering conversation around AI ethics, believes it can be achieved.
“Yes, it can be achieved, but these are necessary conversations between social media owners and information experts. And that can happen the same way it happened years back when mis/disinformation became rampant, that social media platforms had to come up with the 3rd party fact-checking initiative,” he emphasised.
Monsur Hussain, Head of Innovation at CJID, emphasised that prompting Grok to undress women is digital sexual harassment, and it is a clear ethical and legal violation.
He stressed the need for X to take responsibility and safeguard users against this.
“However, the bulk of the fault is not on the users. The big tech responsible for Grok itself has a responsibility to provide safeguards against this. The tool should not be built to accept such user requests,” he added.
Conclusion
While Grok has been helpful to X users, it has also been used in a way that violates women’s privacy and existing ethical principles. Experts fear that if AI ethics are not adopted and Tech giants are not made responsible for AI activities on their platforms, the spread of targeted disinformation generated by AI will continue to negatively impact women.
This report was published under the Dubawa Kwame KariKari Factchecking and OSINT Fellowship.
