A Dangerous Misconception
Among the gravest abuses of generative AI is its use to fabricate sexual imagery of children. A troubling argument sometimes accompanies this conduct: that because no real child was photographed, no real child was harmed. The argument is false, both morally and legally. Synthetic material normalises the sexualisation of children, is frequently built from images of real children taken from social media, fuels demand for abuse, and is increasingly used to groom and to extort victims. It also burdens investigators, who must now distinguish fabricated images from evidence of children in urgent need of rescue.
The Indian Statutory Position
Indian law anticipated this danger with notable foresight. The 2019 amendment to the Protection of Children from Sexual Offences Act, 2012 inserted a definition in Section 2(1)(da) that expressly covers any visual depiction of sexually explicit conduct involving a child, including a digital or computer generated image indistinguishable from an actual child, and images created, adapted or modified so as to appear to depict a child. Synthetic material therefore falls within the Act. Section 14 punishes the use of children for such purposes, and Section 15 penalises storing or possessing such material, whether for transmission, for commercial purposes, or simply by failing to delete, destroy or report it.
Section 67B of the Information Technology Act, 2000 complements this regime. It punishes not only publishing and transmitting, but also creating text or digital images, collecting, seeking, browsing, downloading and promoting material that depicts children in sexually explicit acts. The creation of such content with an AI tool is thus itself an offence.
In Just Rights for Children Alliance v. S. Harish (2024), the Supreme Court held that knowingly viewing, possessing or storing such material without deleting or reporting it is punishable, and urged that the phrase “child pornography” be replaced with child sexual exploitative and abuse material, a term that rightly foregrounds the abuse rather than any notion of consensual content. The 2026 amendments to the IT Rules further compel platforms to label synthetic content and to act against non consensual intimate imagery within hours.
Where the Law Must Go Next
Other jurisdictions are now targeting the tools themselves. The United Kingdom, through its Crime and Policing Act 2026, has introduced offences aimed at AI models designed or adapted to generate such material and at manuals instructing others how to do so. India should consider similar provisions, alongside mandatory safety testing of image generation models, robust content filters, and statutory duties on developers to prevent misuse.
A Shared Responsibility
- Parents should be cautious about posting identifiable images of children publicly.
- Platforms must deploy detection and swift reporting systems, and cooperate fully with law enforcement.
- Any person who encounters such material must not share it, but report it at cybercrime.gov.in, which provides a dedicated channel for crimes against women and children.
A synthetic image may be manufactured, but the violation of childhood it represents is entirely real. The law has already said so with clarity; our task is to enforce it without hesitation.
