Carding
Professional
- Messages
- 2,870
- Reaction score
- 2,493
- Points
- 113
Propaganda is entering a new era. Are you ready?
The project, called CounterCloud, was developed in two months and cost less than $400 a month to maintain. This demonstrates how easy and cheap it is to create mass propaganda.
The creator, who prefers to remain anonymous, used the pseudonym Nea Paw and, according to Wired, is engaged in professional cybersecurity.
In a video posted on YouTube, Paw explained how the project was implemented and why it was created. Paw wanted to see how artificial intelligence-based disinformation would "work in the real world." According to him, the strong language competencies of large language models are ideal for creating and reading fake news articles.
To begin with, Paw entered "opposite" articles in ChatGPT with instructions to write a counter-article. In response, AI offered different versions of the same article, written in different styles and with different points of view.
Paw pointed out that AI creates fake stories, fictional historical events, and casts doubt on the authenticity of the original article.
Next, the "keeper" module was added, thanks to which the AI responded only to the most relevant content. To increase the accuracy, we added audio recordings of news read using AI, as well as photos and profiles of fake journalists.
The system automatically liked and reposted messages corresponding to its narrative, or wrote "counter-tweets" of the opposite content.
In two months, a fully autonomous AI-based system was created, which in 90% of cases generated compelling content around the clock.
At the time of publication, the model was not launched on the Internet. Paw claims: "If you put this project online, it will be impossible to predict where it will go." However, he added that he decided to publish CounterCloud to educate the public about how such systems work from the inside.
The ability of AI to enhance the creation and dissemination of online disinformation has become a major issue in the AI debate. Many expressed concerns about the impact of AI on future elections due to the possibility of creating compelling media materials using personalized information delivery.
The project, called CounterCloud, was developed in two months and cost less than $400 a month to maintain. This demonstrates how easy and cheap it is to create mass propaganda.
The creator, who prefers to remain anonymous, used the pseudonym Nea Paw and, according to Wired, is engaged in professional cybersecurity.
In a video posted on YouTube, Paw explained how the project was implemented and why it was created. Paw wanted to see how artificial intelligence-based disinformation would "work in the real world." According to him, the strong language competencies of large language models are ideal for creating and reading fake news articles.
To begin with, Paw entered "opposite" articles in ChatGPT with instructions to write a counter-article. In response, AI offered different versions of the same article, written in different styles and with different points of view.
Paw pointed out that AI creates fake stories, fictional historical events, and casts doubt on the authenticity of the original article.
Next, the "keeper" module was added, thanks to which the AI responded only to the most relevant content. To increase the accuracy, we added audio recordings of news read using AI, as well as photos and profiles of fake journalists.
The system automatically liked and reposted messages corresponding to its narrative, or wrote "counter-tweets" of the opposite content.
In two months, a fully autonomous AI-based system was created, which in 90% of cases generated compelling content around the clock.
At the time of publication, the model was not launched on the Internet. Paw claims: "If you put this project online, it will be impossible to predict where it will go." However, he added that he decided to publish CounterCloud to educate the public about how such systems work from the inside.
The ability of AI to enhance the creation and dissemination of online disinformation has become a major issue in the AI debate. Many expressed concerns about the impact of AI on future elections due to the possibility of creating compelling media materials using personalized information delivery.
