Artificial Images, Real Harm: Does Indian Law Adequately Address AI-Generated CSEAM?
~ Riya Bhandari
ABSTRACT
The rise of generative artificial intelligence has created new challenges in combating online child sexual exploitation. AI-generated Child Sexual Exploitative and Abuse Material (CSEAM) complicates traditional legal frameworks by enabling the creation of sexually explicit depictions of children without necessarily involving direct physical abuse. This article examines whether India’s existing legal framework is adequately equipped to address AI-generated CSEAM and the challenges that continue to persist in this rapidly evolving digital landscape. It analyses the Supreme Court’s decision in Just Rights for Children Alliance v. S. Harish (hereinafter “JRCA”), which extended the scope of child protection laws to cover AI-generated material that appears to depict children. The article argues that the judgment adopts a progressive, harm-based approach by recognising the broader societal and psychological harms associated with such content. However, it also highlights unresolved concerns relating to the appearance test, constructive possession, and evidentiary challenges in an increasingly complex digital environment.
Keywords: Artificial Intelligence, Child Sexual Exploitative and Abuse Material (CSEAM), Child Protection, Deepfakes, POCSO Act
INTRODUCTION
The emergence of Artificial Intelligence has transformed child sexual exploitation into a rapidly evolving digital challenge. Generative AI tools can now create highly realistic images and videos with minimal effort, blurring the line between reality and fabrication. While these technologies offer significant benefits, they have also created new avenues for abuse. Among the most disturbing developments is the rise of AI-generated Child Sexual Exploitative and Abuse Material (CSEAM), which can depict children in sexually explicit situations with alarming realism, including children who do not exist in reality.
This development raises a difficult legal question: if no real child was involved in the creation of such material, can it still amount to child sexual exploitation? As AI-generated content becomes increasingly sophisticated and accessible, legal systems across the world are grappling with how existing child protection laws should respond.
In India, the Supreme Court’s decision in JRCA marked a significant step towards strengthening protections against online child sexual exploitation through a child-centric approach.
AI-GENERATED CSEAM – A NEW FRONTIER OF CHILD EXPLOITATION
The rapid advancement of generative Artificial Intelligence has transformed the landscape of online child sexual exploitation. AI tools can now create highly realistic images and videos, making it increasingly difficult to distinguish fabricated content from reality. While these technologies offer remarkable opportunities for innovation, they have also created new avenues for abuse. A dangerous misconception is that AI-generated CSEAM is a victimless crime because no child is physically involved in its creation. However, the absence of physical abuse during production does not eliminate the harm. AI-generated sexual imagery can cause profound psychological trauma, particularly when real children’s images are manipulated through deepfakes. Victims may experience digital revictimization where exploitative content repeatedly prolongs and reproduces the trauma associated with abuse, loss of dignity, reputational harm, and the continued fear that abusive content may be viewed, shared, or recreated indefinitely.
If no real child is visible in an AI-generated image, should the law still be concerned? AI-generated CSEAM does not exist in a vacuum. It normalises child sexual exploitation, encourages demand for abusive content, and enables the creation of deepfake imagery targeting real children. The growing evidence suggests that it should. Criminological studies indicate that artificial abuse material may reduce inhibitions, intensify offending behaviour, and encourage the consumption of actual child sexual abuse material. AI-generated content also creates operational harms by overwhelming investigative systems, obscuring genuine victims within vast datasets, and diverting resources away from children requiring immediate intervention.
The scale of the threat is already alarming. In 2025 alone, the CyberTipline received 21.3 million reports containing more than 61.8 million images, videos and other files relating to suspected child sexual exploitation. Similarly, INHOPE member hotlines processed almost 2.5 million suspected online records of child sexual abuse and exploitation material. Recent reports demonstrate that this threat is no longer theoretical. Researchers at Stanford Internet Observatory and Thorn found that open-source AI image-generation tools were already being misused to create child sexual abuse material, including customised images of specific victims. The Internet Watch Foundation (IWF) has also reported realistic AI-generated abuse videos using deepfake technology and warned that future iterations of AI will become even more realistic and difficult to detect. Its 2024 report highlights the growing presence of AI-created CSEAM across both the open web and dark web, underscoring the urgent need for an effective legal response.
INDIA’S LEGAL RESPONSE TO AI-GENERATED CSEAM
India’s legal response to online child sexual exploitation is primarily anchored in the Protection of Children from Sexual Offences (POCSO) Act, 2012 and Section 67B of the Information Technology Act, 2000. Together, these provisions criminalise the creation, possession, transmission and consumption of child sexual abuse material in digital spaces. However, when it comes to generative artificial intelligence and its emerging forms of exploitation, a significant development came with the Court’s decision in JRCA. At a time when technological advancements threatened to outpace existing legal protections, the Court adopted a purposive and child-centric interpretation of the law. Emphasising that the protection of a child’s dignity is the “spine” of POCSO, it refused to allow digital technicalities to undermine child protection.
The Court clarified that liability under Sections 13 to 15 of POCSO Act extends beyond the creation of exploitative material to its possession and storage. By characterising these offences as inchoate offences, the Court recognised that the harm posed by CSEAM often begins long before its actual distribution. This approach shifts the focus from completed acts of exploitation to the prevention of exploitation itself.
Most importantly, the judgment brought AI-generated content within the ambit of child protection laws. Interpreting Section 2(1)(d) of POCSO Act, the Court held that material may constitute CSEAM even where no identifiable child exists, provided it is indistinguishable from, or appears to depict, a child. In doing so, the Court rejected a narrow victim-based approach and acknowledged that digital representations can perpetuate exploitation, normalise abuse and fuel demand for child sexual abuse material. The judgment therefore represents one of India’s most significant judicial responses to the challenges posed by AI-enabled child exploitation.
INTERDISCIPLINARY ANALYSIS
Compared to India, several jurisdictions have begun addressing AI-generated CSEAM through more explicit legislative measures. The United Kingdom already criminalises both photorealistic and non-photorealistic child abuse imagery and is moving towards regulating AI tools specifically designed to generate such content. Several states in the United States have similarly enacted laws expressly covering AI-generated CSEAM, while the European Union is actively considering reforms that would directly address AI-enabled child exploitation. Unlike these jurisdictions, India’s response remains largely judge-made. The Supreme Court’s decision in JRCA undoubtedly strengthens child protection, but it also raises questions that may require legislative intervention.
DOES THE LAW FULLY SOLVE THE AI PROBLEM?
One major such concern is the Court’s reliance on the “appearance test”. By allowing judges to determine whether material depicts a child from the perspective of an ordinary prudent person, the Court avoided the practical difficulty of requiring forensic age verification in every case. However, the standard remains inherently subjective. As AI-generated imagery becomes increasingly realistic, the distinction between a child and a young-looking adult may not always be obvious. The absence of a clearer benchmark could therefore lead to inconsistent outcomes across cases.
The judgment also expands the concept of possession through the doctrine of constructive possession. A person may be considered in possession of CSEAM even without downloading or storing it, provided they exercise control over the material while viewing it online. While this prevents offenders from exploiting technological loopholes, the Court’s reference to a “reasonable window” leaves room for uncertainty. How long is reasonable? At what point does accidental viewing become knowing possession? Greater clarity on these questions would strengthen legal certainty without undermining child protection. The operation of Section 30 of POCSO Act presents a further challenge. Once foundational facts are established, the burden shifts to the accused to rebut the presumption of a culpable mental state. This undoubtedly strengthens enforcement in an area where evidence is often difficult to obtain. Yet, in an increasingly complex digital environment, situations involving malware, automatic downloads, shared devices, or unauthorised access may complicate the assessment of intent. While the Court recognised that such issues must be examined during trial, proving the absence of knowledge or intent may remain a difficult task for many accused persons.
The judgment therefore marks an important step forward. However, as AI technologies continue to evolve, questions of consistency, digital ambiguity, and procedural fairness suggest that judicial interpretation alone may not fully resolve the challenges posed by AI-generated CSEAM.
WAY FORWARD
The rapid advancement of artificial intelligence demands an equally responsive legal framework. While the Supreme Court has significantly strengthened India’s approach to AI-generated CSEAM through judicial interpretation, long-term certainty requires legislative action. Explicit statutory provisions addressing AI-generated CSEAM would eliminate ambiguity and provide clearer guidance for enforcement. Greater judicial clarity on the application of the appearance test would also help ensure consistency across cases involving increasingly realistic digital content. At the same time, specialised cyber-investigation mechanisms and enhanced digital forensic capabilities are essential to effectively investigate AI-related offences and navigate the technical complexities they present. Such reforms would reinforce child protection while ensuring that the law remains equipped to meet emerging technological challenges.
CONCLUSION
The debate surrounding AI-generated CSEAM is no longer about the future, it is a challenge that has already arrived. By recognising that digital abuse can inflict real harm even where no physical child is directly involved, the Supreme Court has taken an important step towards ensuring that technology does not outpace child protection laws. Yet, as artificial intelligence becomes more sophisticated, the law must move beyond merely keeping pace and begin anticipating emerging threats. The true test of any child protection framework lies not in how it responds to yesterday’s forms of exploitation, but in how effectively it confronts those of tomorrow. In an age where abuse can be created with a prompt and shared in seconds, safeguarding children’s dignity cannot remain a reactive exercise, it must become a proactive legal imperative.
Author’s Bio – Riya Bhandari is a second year B.B.A LL. B (Hons.) student at National Law University, Nagpur. She can be reached at riyabhandari1802@gmail.com
REFERENCES
- Just Rights For Children Alliance vs Harish, 2024 INSC 716
- United Nations Children’s Fund(UNCIEF), Childhood In A Digital World(2025), <https://www.unicef.org/innocenti/media/11296/file/UNICEF-Innocenti-Childhood-in-a-Digital%20World-report-2025.pdf>
- Internet Watch Foundation (IWF), AI CSAM Report 2026: Harm Without Limits (2026), <https://www.iwf.org.uk/media/hl1nvdti/iwf-ai-csam-report-2026.pdf>
- Parth Malhotra, Artificial Intelligence Generated Child Sexual Abuse Material (CSAM): A Legal View, Indian Journal of Law and Legal Research, Vol. VII Issue IV, ISSN 2582-8878.
- National Center for Missing & Exploited Children (NCMEC), CyberTipline Reports and Data, <https://report.cybertip.org/>
- United Nations Children’s Fund (UNICEF), Deepfake Abuse Is Abuse, UNICEF Press Release (2024), <https://www.unicef.org/press-releases/deepfake-abuse-is-abuse>