С 17 августа 2026 года в Сингапуре создание интимных изображений — в том числе полностью синтетических, сгенерированных искусственным интеллектом — без согласия будет квалифицироваться как отдельное уголовное преступление. Тот же пакет поправок также ...
Сообщение Planet-Today.com. Перевод заголовка и краткого описания выполнен автоматически.
From 17 August 2026, Singapore makes the production of intimate images — including fully synthetic AI-generated ones — without consent a specific criminal offence. The same package of amendments also strengthens penalties for sexual offences involving minors and overseas grooming, while adjusting the use of judicial caning for certain other crimes. Officials frame the changes as closing technological gaps; critics of broad digital regulation elsewhere watch for enforcement realities.
 |
| Source: Magnific |
Key Takeaways by Planet Today
Core new offence: Producing intimate images, including completely AI-generated material, without the depicted person’s consent becomes punishable by up to two years’ imprisonment, a fine, or both.
Child protection focus: If the material depicts a minor under 14, imprisonment of up to two years is mandatory and the offender may also face a fine or caning.
Extraterritorial reach: Authorities can prosecute grooming of minors abroad followed by sexual offences when either the perpetrator or the victim travelled from Singapore before the crime.
Caning adjustments: Judicial caning remains available for serious sexual offences; for several non-sexual crimes (including certain forms of piracy, attempted robbery and carrying illegal weapons) it shifts from mandatory to discretionary.
Policy logic: The government argues that pure synthetic images are “no less objectionable” than altered real ones and that clearer rules reduce both enforcement barriers and the supply of abusive material.
Singapore’s Ministry of Home Affairs confirmed on 14 August 2026 that key provisions of the Criminal Law (Miscellaneous Amendments) Act 2025 take effect on Monday, 17 August. The centrepiece for digital harms is a new offence of producing intimate images without consent. The definition of “intimate image” is explicitly expanded to cover material generated by artificial intelligence that depicts a person without any alteration of an existing photograph or recording.
In official language, “the increasing sophistication of AI models means that completely synthetic images can be generated. These are no less objectionable than images which are created by altering existing images or recordings.” Offenders face up to two years in prison, a fine, or both. Where the image depicts a person under 14, the court must impose imprisonment of up to two years and may also order a fine or caning.
Parallel changes tighten the treatment of sexual offences against minors and close a previous gap in jurisdiction over grooming that occurs partly outside Singapore. If either the offender or the victim left the city-state before the offence, prosecution becomes possible. For victims under 14 the maximum sentence for such overseas-linked offences rises to seven years’ imprisonment.
What the Law Actually Changes
Earlier Penal Code provisions already criminalised possession of and access to intimate images and most forms of deepfake material created by altering real images. The 2025 amendments add three clarifying steps: (1) pure synthetic AI images fall inside the definition of intimate images; (2) the act of production itself is expressly criminalised even without further distribution; and (3) computer-generated child abuse material is covered even when prosecutors cannot prove that an image of a real child was used in its creation. The policy intent, stated repeatedly by the Ministry of Home Affairs, is to remove technical arguments that pure AI output falls outside existing statutes.
Senior Minister of State Sim Ann told Parliament during the bill’s passage that the measures aim “both to reduce the roadblocks to enforcement and to reduce the supply of such (AI-generated) materials, which contributes to a vicious cycle of abuse.” The same package leaves judicial caning intact for serious sexual crimes such as rape and aggravated molestation while converting mandatory caning into discretionary caning for a list of other offences that includes piracy under the law of nations, attempted robbery, and certain weapons offences.
Context and Parallel Measures
The criminal amendments sit alongside the Online Safety (Relief and Accountability) Act 2025, whose Online Safety Commission began operations earlier in 2026. That body gives victims of intimate-image abuse, image-based child abuse, doxxing and online harassment a dedicated channel for rapid platform takedown requests and civil remedies. Together the two frameworks create both a criminal deterrent and a civil relief pathway.
Globally, non-consensual intimate imagery — whether traditional “revenge porn” or AI-generated — has become a documented harm, with studies cited by Singaporean ministers indicating that the large majority of deepfake videos circulating online involve non-consensual sexual content, disproportionately targeting women. Singapore’s approach treats pure synthetic images of real or realistic-looking persons as functionally equivalent to altered photographs for the purposes of harm and criminal liability.
Alternative Perspectives and Open Questions
Supporters of the law, including government speakers and many child-protection advocates, argue that technology-neutral or technology-updating criminal rules are necessary once generative tools make high-quality non-consensual imagery trivial to produce at scale. They point to the difficulty of proving that an image was “altered” once models can invent photorealistic faces and bodies from text prompts alone.
A quieter set of questions appears in legal and technology commentary elsewhere. One concerns practical enforcement: identifying the producer of a purely synthetic image that never left a private device, or attributing creation when tools are open-source and run locally. Another concerns the boundary between criminal production and legitimate artistic, satirical or research uses of synthetic imagery — although Singapore’s statute is framed around intimate images of persons and consent, not around all generative content. A third, raised in comparative discussions of UK, South Korean and Australian approaches, is whether equal severity for real-child and fully synthetic child-like material is the optimal calibration, or whether the absence of a real victim should affect sentencing ranges. Singapore’s Ministry of Home Affairs has stated it does not intend to create lighter penalties for AI-generated child abuse material; the resemblance to a real child is sufficient.
Public reaction inside Singapore has been limited in the immediate reporting cycle. Prior parliamentary questions from opposition MPs focused on police practice when victims report AI-generated intimate images circulating on adult websites, and on whether such cases are treated as arrestable Penal Code offences or steered toward civil Protection from Harassment Act routes. Official replies emphasised case-by-case investigation rather than a default non-criminal pathway.
Two Short Additional Notes
First, the same legislative package that expands digital sexual-offence tools also recalibrates caning for a set of older, non-sexual crimes, moving several from mandatory to discretionary. The government presents this as proportionality review rather than abolition. Second, the Online Safety Commission’s parallel civil track means victims of intimate-image abuse now have both a criminal complaint route and a faster administrative takedown and damages route — a dual structure that is still relatively rare internationally.
Readers following digital regulation more broadly may find related discussion of platform accountability and online harms useful in wider technology-policy coverage.
Where the Rules Now Stand
As of 17 August 2026, creating a non-consensual intimate image of another person with generative AI is a free-standing criminal offence in Singapore, carrying the same baseline penalties as voyeurism and possession of intimate images. Aggravated rules apply when the depicted person is under 14. Overseas grooming linked to Singapore by travel of either party is more clearly prosecutable and more heavily punishable when the victim is young. Caning remains a sentencing option for the most serious sexual crimes and becomes discretionary for a narrower list of other offences.
The legislation is deliberately technology-updating rather than technology-specific. Its long-term test will be whether the new production offence is used, how often pure synthetic cases reach court, and whether the combination of criminal liability and civil relief reduces the volume of non-consensual intimate imagery circulating among Singapore users. Those outcomes will be measured in future enforcement statistics, not in the text of the statute itself.
Primary sources: Ministry of Home Affairs statements and the Criminal Law (Miscellaneous Amendments) Act 2025; AsiaOne and AFP reporting of 14 August 2026 confirming the 17 August commencement; parliamentary remarks by Senior Minister of State Sim Ann; MHA replies on AI-generated material and child-abuse offences. Official MHA pages on the Bill’s first reading and commencement notifications provide the legislative text and rationale.
Disclaimer: This article summarises enacted amendments and official explanations current as of 14 August 2026. It is not legal advice. Individual cases turn on specific facts; readers should consult the Singapore Statutes Online text and qualified counsel for definitive interpretation. Enforcement practice may evolve after the provisions take effect.
