A recent report by The Bureau of Investigative Journalism (TBIJ) showed how a Sri Lankan influencer has been populating Facebook pages with racist, xenophobic, and anti immigrant AI generated content and making hundreds of thousands dollars in the process.
Geeth Sooriyapura, labelled “The King of AI slop” by TBIJ but also known by his preferred moniker of “Facebook Ads king”, claims to have made $300,000. He is also now running classes on how to use these tools to rake in cash using similar methods.
Their posts have included claims that a new housing scheme by London Mayor Sadiq Khan’s will be exclusively provided to Muslims. Another, described by TBIJ, featured an AI generated video which:
“[…] shows a large naval ship colliding with two inflatable dinghies full of people – representing those crossing the Channel to seek asylum in the UK – who are then thrown out of the boats and into the sea. The caption reads: “Who really wants to see something like this?”
Sooriyapura is an example of what I have called “locally recruited agents of influence”. These are a category of online actors who enable the spread of disinformation and a particular form of propaganda using a technique of “reflexive control”.
Locally recruited agents of influence are working in the interests of particular political actors who may want a particular political outcome (e.g. in an election) or just want to sow division and discord. But they are only very loosely (if at all) connected to one another.
People like “The King of AI Slop” most likely are not connected to anyone who wants to promote anti-immigrant sentiment for political reasons and might even disagree with, or be indifferent to, such ideas themselves. But because they have identified a way to make money from promoting these ideas they have nevertheless become part of the ecosystem which enables the spread of disinformation.
He stated that he targets older British people because they don’t like immigrants, are highly politically engaged and easily triggered to engage with content. New AI tools which can quickly generate image and video content has made it easier than ever to stimulate emotional responses. This turns the audiences he targets into “unwitting agents of influence” who also help in the spread of toxic messages by watching and sharing Sooriyapura’s content.
Sooriyapura makes money from Meta who will pay a share of revenue from ads which play before, during or after viral videos and from selling subscriptions to his pages. While he callously uses this system to manipulate and spread hatred it is the design of the platform and it’s revenue model which enables and incentivizes this.
The case of the King of AI Slop shows how AI tools are making it easier and more lucrative than ever to draw more people into systems which manipulate people through disinformation. While he, and those he teaches, might make some decent money from their activities they are still dancing to someone else’s tune and bringing others along with them.
This is especially dangerous when the US government is openly expressing the same kinds of views as those pushed by financially motivated actors like Sooriyapura warning against “civilizational erasure” due to non-European immigration into the continent (rhetoric which Russia has supported).
You can read my Open Access paper on “locally recruited agents of influence” and “reflexive control” or a shorter blog post version.