Sri Lankan network uses AI to monetise anti-migrant narratives in the UK

A network run from Sri Lanka is creating and monetising anti-immigrant, Islamophobic content targeting UK audiences using AI-generated posts, stoking division while raking in substantial ad revenue. What happened A Sri Lankan influencer, Geeth Sooriyapura, is linked to a coordinated network of 128 Facebook pages and groups focused on anti-migrant and anti-Muslim content aimed at British users, a ccording to a joint investigation by the Institute for Strategic Dialogue (ISD) and the Bureau of Investigative Journalism . The network is estimated to have around 1.6 million followers across its pages. Much of the content is AI-generated imagery and video (e.g., fake mass protests, deep-fake style clips), used to amplify divisive narratives. Examples include: a video of a naval ship hitting dinghies full of asylum seekers (AI-generated), and false claims that London council housing is reserved for Muslims. Sooriyapura runs an “online academy” teaching others how to build similar Facebook pages, monetise them via Meta’s tools (in-stream ads, subscriptions, performance bonuses), and use generative AI, including ChatGPT, to generate ideas and content, and scale content fast. He claims to have earned USD 300,000 from the operation. Some of the pages push anti-Labour party messaging (e.g., targeting UK Prime Minister Keir Starmer), Islamophobic conspiracy theories (e.g., “replacement” narratives), and portray migrants as invaders. Meta has removed some of these pages for “inauthentic behavior.” Why it happened This happened because social media monetisation models, coupled with weak governnace, transparency and accountability from major platforms such as Meta, allowed Sooriyapura's network to profit from hate and misinformation with little or no oversight. Sooriyapura's operation appears to be purely profit-driven, not overtly ideological. He openly instructs students to monetise engagement by tapping into polarising political themes Algorithmically promoted engagement - regardless of truth or harm - created a feedback loop: inflammatory AI-generated content received more views and shares, leading to more ad revenue, while the origins and coordinated nature of these campaigns remained obscured from users and sometimes even from regulators. The influencer openly taught these tactics and boasted of targeting “old people” in the UK, a demographic he claimed was most susceptible to these narratives. He also took advantage of the ease of use and scalability of generative AI tools, which Sooriyapura and his network used to create emotionally provocative content fast, reducing barriers to entry and increasing the volume of output. In addition, Meta’s content moderation, including the labelling of AI-generated content and disinformation, is weak and inconsistent, allowing these kinds of pages to exist and grow. What it means For UK society : Such content deepens social divides by stoking fear, xenophobia, and Islamophobia. It may fuel real-world prejudice and hostility toward migrants and Muslim communities. For political discourse : The disinformation network injects false narratives into UK politics, potentially influencing public perception and even engagement by mainstream political actors. For Meta/platforms : Raises serious questions about how effectively Meta enforces its policies on AI content, political advertising, and monetisation of hateful or polarising speech. For other actors : Signals a worrying trend that commercially motivated disinformation campaigns, powered by AI, are not just the purview of state actors but also private individuals exploiting platforms for profit. For broader society : Undermines trust in online information, encourages polarisation, and could contribute to radicalisation or normalisation of hate speech. System 🤖 ChatGPT Developer: OpenAI Country: UK Sector: Politics Purpose: Monetise controversial content Technology: Generative AI Issue: Auth enticity/integrity; Mis/disinformation; Transparency Research, a

Date it happened

2025-01-01

Organisation involved

Geeth Sooriyapura

Product, system or model

ChatGPT

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.