myST+MenuChoose edition

Sign up now: Get ST's newsletters delivered to your inbox

AI deepfake nudes rarely exist in isolation, often triggering harassment, reputational damage and psychological distress.

SummaryGrok AI controversy allowed creation of deepfake nudes, causing harassment and distress to victims.Singapore amended laws to criminalise deepfake sexual material and introduced OSRA for content removal.Social attitudes normalising sexualised manipulation exacerbate deepfake abuseAI generated

Published Feb 09, 2026, 05:00 AM

Updated Feb 09, 2026, 05:00 AM

SINGAPORE – In August 2023, Singaporean content creator Mathilda Huang, 28, found deepfake nudes of her on “seedy websites”.

The person who posted the pictures falsely claimed that they came from her OnlyFans account, a subscription-based platform where creators share adult content with paying fans.

Ms Huang recounted having to go through the “humiliating” task of tracking down websites to request their removal from search results.

In recent times, the proliferation of “AI nudifying tools” – online services that use artificial intelligence (AI) to digitally strip subjects – on forums, Telegram groups and third-party app stores has worsened the problem.

The problem reached new heights in December 2025 when Elon Musk-backed social media platform X allowed its built-in AI chatbot Grok to put women and

children in transparent bikinis or portray them in provocative ways.

Regulators in the United Kingdom, European Union and France launched investigations into Grok, while governments in Indonesia, Malaysia and the Philippines blocked access to the tool

AI deepfake nudes often trigger real-world harassment, reputational damage and psychological distress. Governments acted promptly, and rightly so, to protect their citizens.

For one thing, Ms Huang was harassed after her fake nude images were circulated online.

“There were a slew of direct messages that came in on my social media, which asked me how much I was charging for a night and solicited me for sexual services,” she told The Straits Times.

She avoided people and late nights. “It felt like my inner world was too heavy,” said Ms Huang, who still does not know who created the deepfake nudes.

The sense of betrayal and humiliation deepens when deepfakes of victims are shared among people they know, inviting real-life dangers of bullying, sextortion, intimidation and stalking.

A case in point: A report of extortion was filed after the deepfake nudes of over 20 teenage girls in the Spanish town of Almendralejo were created and circulated by their schoolmates in September 2023.

Researchers from Campus Sexual Misconduct in a Digital Age at the National University of Singapore pointed out that such proximity to perpetrators is damaging. Lead researcher Michelle Ho said: “This makes the harm feel inescapable and deeply personal.”

Non-profit organisation SG Her Empowerment’s chief executive officer How Kay Lii, whose organisation supports deepfake victims, said: “Many victims spoke about panic, fear that the content would resurface, sleeplessness, anxiety and, in some cases, thoughts of self-harm.”

Although headline-grabbing, a blanket ban may not effectively safeguard citizens, said experts.

Institute of Policy Studies senior research fellow Chew Han Ei said: “Evidence so far suggests bans often displace harm with users switching platforms or using workarounds.”

That was what happened in Malaysia and Indonesia despite the widely publicised bans.

Users in Malaysia and Indonesia reportedly still had access to Grok – which can be downloaded as an app, and connected via a browser or within X, which was not banned. Users can also bypass the ban by using a virtual private network to mask their locations.

Singapore has chosen to engage X on the issue for now,

one-stop agency, the Online Safety Commission (OSC),

For one thing, the OSC will be able to issue takedown orders for deepfake images, which is fundamentally what most victims want.

New legislation passed in Parliament in November 2025 – the Online Safety (Relief and Accountability) Act

Mr Gilbert Leong, a senior partner and head of Dentons Rodyk’s intellectual property and technology department, said: “Singapore’s approach strives to give an individual the ability to get recourse, while a blanket ban would not.”

Societal tolerance for deepfake nudes must also be lowered to truly eradicate the harms. Ms Sugidha Nithiananthan, director of advocacy and research at women’s rights group Aware, said: “When society downplays these harms, it compounds survivors’ trauma, making them feel invalidated and isolated.”

Li Ying is a technology correspondent covering online safety, cybersecurity and the impact of artificial intelligence.

AI/artificial intelligence

Need help? Reach us here.