.jpeg)
Each week, a new horror surfaces from the excavation of the Chemmani mass grave in Jaffna. From the skeletons of infants buried with their mothers, to school bags and children’s toys, each discovery underpins years of allegations that the area, largely occupied by the Sri Lankan army, served as a site of disappearances and torture. Local media coverage of the excavation has sparked widespread outrage amidst Tamil society, and protests calling for justice have gathered momentum, largely thanks to the work of local photojournalists such as Kumanan Kanapathipillai and Prabhakaran Dilakshan who have meticulously documented the discoveries.
While the atrocities of Chemmani may have occurred decades ago, those seeking justice have had to contend with a phenomenon entirely of the current moment, namely the proliferation of AI-generated imagery of the mass graves. As the excavation unfolded, social media users shared AI-generated images depicting the most distressing discoveries in more graphic ways, from exaggerating the images of skeletons as fleshed out and clothed bodies, to simple recreations of the existing photography with - to the trained eye - obvious distortions. Eventually, social media posts about Chemmani became so oversaturated with AI, lawyer Ranitha Gnanarajah, representing victim families, warned that AI-generated images of the human remains risked distorting the criminal case around the mass grave. Ms Gnanarajah said that those sharing such images may even be at risk of contempt of court charges.
As well as the threats that fake images can present to the integrity of ongoing legal proceedings, the moral pitfalls of generative AI in the context of justice for atrocities must be grappled with. These questions first gained prominence in the context of the genocide in Gaza, when an AI-generated image with the slogan ‘All Eyes on Rafah’ was shared on Instagram in May 2024. The post gained unprecedented virality, being shared by more than 47 million users, and was seen as an easy action people could take to express their opposition to Israel’s bombardment of Gaza. While the post was criticised by some for failing to show the realities of the genocide and lacking a meaningful call to action, the most severe consequence has been the opening of the AI floodgates in content related to Gaza and Palestine. Despite the near daily broadcast of Israel’s atrocities in Gaza, documented both by Palestinian journalists and ordinary civilians, the algorithm for many social media users will prioritise the AI ‘slop’ capitalising on keywords and hashtags. As well as giving rise to accusations of inauthenticity across the board with atrocity reports, such content risks creating apathy among onlookers, while often real footage is censored by social media platforms.
Like Gaza, many aspects of the Tamil genocide were recorded. LTTE videography, civilian phone footage, the work of Tamil journalists and Sri Lankan army trophy videos all serve as documentation of the atrocities enacted upon the Tamil people by the Sri Lankan state. They required no manipulation to reveal the devastating and brutal realities of the genocide. Sixteen years later, we continue to unearth atrocities likely spanning much further back than the apocalyptic months of Mullivaikkal. The photographs from Chemmani, largely of brown earth and grey human remains, punctured by the vibrant blue of a UNICEF school bag and the bright colours of a child’s doll, are devastating enough in their stark reality. Distorting and sensationalising those images with generative AI dilutes the documentation of these atrocities and debases the humanity of the victims, for the sake of likes and clicks.
Although we may live in the social media age, virality does not equate to justice. Above all, the most important records we have are the testimonials of those who survived and the silences of those who did not. We can hope that with forensic advances, those discovered at Chemmani can be identified and laid to rest with dignity. However, we cannot reanimate them, especially not with AI.