31 JANUARY, 2025
Who: Charles Darwin University forensic linguistics expert Dr Awni Etaywe is a lecturer in linguistics and a forensic linguistics researcher focusing on terrorism, incitement to hatred, radicalisation and genocide, and digital deviance.
Dr Etaywe fed sections from a far-right, white supremacist violent extremist manifesto to DeepSeek and ChatGPT and compared and contrasted their effectiveness in flagging disinformation tactics.
Contact details: Call +61 8 8946 6721 or email [email protected] to arrange an interview.
Quotes attributable to Dr Etaywe:
“Although some may see both ChatGPT and DeepSeek in a kind of AI race between the US and China, the science behind both present both platforms to researchers as tools that can aid in flagging and combatting disinformation, each complementing the other with the multifaceted analytical nuances they offer.
“ChatGPT provides a broad, detailed analysis that helps users understand the structural and rhetorical components of extremist discourse. Its focus on linguistic techniques and logical fallacies gives a deep, analytical framework for debunking harmful narratives.
“On the other hand, DeepSeek excels in identifying the emotional and psychological manipulations that underlie disinformation, making it an essential tool for understanding how extremist rhetoric seeks to radicalise individuals on a more visceral level.
“ChatGPT helped break down how false narratives are constructed, focusing on rhetorical techniques such as false authority, exaggerated statistics, and selective presentation of facts. By evaluating the text through the guiding questions, ChatGPT dissected how extremist ideology is spread, exposing the misleading claims and manipulative framing that contribute to the radicalisation process.
“ChatGPT identified key strategies such as the distortion of demographic trends, the misrepresentation of immigration, and the cherry-picking use of historical figures’ quotes to support extremist claims. False framing was found, highlighting how Wikipedia is misused to lend legitimacy to the text’s baseless claims.
“While also focusing on identifying disinformation techniques, DeepSeek emphasises the emotional appeals and the creation of a divisive narrative, which is central to radicalising audiences. DeepSeek’s questions delve into how extremist views are reinforced through fabricated stories, cherry-picked evidence, and inflammatory language. It underscores the manipulation of fear, identity, and urgency as key drivers of radical ideologies.
“Together, DeepSeek and ChatGPT not only support debunking efforts but also empower individuals, organisations, and governments to develop more informed counter-narratives that can help undermine the effectiveness of extremist ideologies. Both platforms contribute to the growing body of AI-powered solutions that are essential in the ongoing battle to preserve truth, promote social cohesion, and counter the growing influence of disinformation in the digital age.”
Contact details:
Raphaella Saroukos she/her
Research Communications Officer
Marketing, Media & Communications
Larrakia Country
T: +61 8 8946 6721
E: [email protected]
W: cdu.edu.au