A cybersecurity professional who chose to remain anonymous, with the name Nea Paw, has built an AI-driven propaganda machine in less than two months. The project, named “CounterCloud,” demonstrates the ease and affordability of creating mass propaganda campaigns in the digital age.
Powered by OpenAI’s cutting-edge tools, notably ChatGPT, CounterCloud operated at a monthly cost of less than $400. Paw revealed that the goal of the project was to show how AI-driven disinformation can be used to manipulate public opinion.
To create CounterCloud, Paw first fed contradictory articles into ChatGPT to get the AI to generate opposing narratives. The AI then produced a plethora of divergent versions of the same article, each with its own unique style and perspective.
To make CounterCloud more credible, Paw added a gatekeeper module that filtered responses for relevance. Audio clips of AI-generated articles being narrated by simulated newsreaders and counterfeit journalist profiles were also used to lend an air of authenticity.
Finally, Paw programmed CounterCloud to promote content that aligned with its agenda and to generate “counter” tweets against opposing views. After two months, CounterCloud had evolved into a fully autonomous AI system that could generate “convincing content” 90 per cent of the time, 24/7.
The sources for this piece include an article in BusinessInsider.