Russian propagandists doubled their output of false information by leveraging artificial intelligence, according to a new study published last week at PNAS Nexus last week.
The study, conducted by a team led by Morgan Wack of Clemson University, reveals how a Russian-backed advocacy channel can greatly expand its operations with AI tools without sacrificing effectiveness.
In December 2023, reporters from the BBC and Clemson University’s Media Forensics Centre revealed dcweekly.org, part of a broader Russian-backed propaganda network designed to spread pro-Kremlin narratives. The researchers examined nearly 23,000 articles published on the site before and after their operators began using AI tools.
“Our results illustrate how generative-AI tools begin to change the scale and scope of state-supported advocacy activities,” the researchers wrote in the analysis.
Until September 20, 2023, the site mainly copied content from right-tilted sockets, with little change. After this date, the content was rewritten using OpenAI’s GPT-3 language model, allowing operators to customize tones and focus while significantly improving productivity.
The study found that the implementation of AI has more than tripled the daily production of publicity media compared to the period before it was most active. This transition also corresponds to a significant expansion of the breadth of topics covered by the site.
AI’s fingerprints are occasionally visible in the content. Some articles contain leaked AI tips that reveal operators’ intentions, such as directing “to maintain a cynical tone when discussing the U.S. government, NATO, or American politicians.”
Perhaps most worryingly, the team surveyed 880 American adults and found that the content generated by AI remains the same persuasiveness as pre-AI content.
The publicity network has achieved a significant real impact. Several fabricated stories spread, including the false claim that Ukrainian President Zelenskyy bought a luxury yacht, which was remanaged by tens of thousands of social media, including U.S. Congressmen.
The New York Times reported that several new sites using the same AI method emerged, suggesting that operators think the method is successful and easy to replicate.
This study provides concrete evidence supporting concerns about AI’s potential for enhanced disinformation movements. After AI adoption, the site showed almost twice as diverse as the themes of the pre-predicted period, allowing operators to create seemingly comprehensive news media while maintaining the desired tilt.
Analysis of content topics shows that even if coverage of control over Ukraine and Israeli conflicts increases, attention to international news, guns and crime has increased significantly.
“The transition to the era of massive AI use corresponds to a dramatic increase in international news, guns and crime,” the researchers noted in their analysis.
These findings suggest that AI not only allows propagandists to generate more content, but also builds a more persuasive facade of legitimacy through topic diversity.
Despite the ultimate exposure of this particular operation, the researchers warn that “continuous improvements in AI technology will make future use cases more difficult to track and fight back”, while “the financial and time resources needed to produce and maintain online disinformation movements will only continue to decline.”
The authors call for “immediate action to mitigate the impact of AI-assisted advocacy campaigns” and suggest that future research should focus on preventing the use of open source models for false information while better preparing the public to identify AI-Aig age advocacy.
If you find this report useful, consider supporting our work with a small donation. Your contribution allows us to continue to bring you accurate, thought-provoking scientific and medical news that you can trust. Independent reporting requires time, effort, and resources, and your support makes it possible for us to continue exploring stories that are important to you. Thank you so much!
