“I think digital platforms need to open us to researchers like us, because what they have is data. In order for us to train, in order for us to understand how things work at a massive scale, they would have to have a handshake with us.” Dipto Barman
“I think digital platforms need to open us to researchers like us, because what they have is data. In order for us to train, in order for us to understand how things work at a massive scale, they would have to have a handshake with us.”
Dipto Barman
In the latest episode of the ADAPT Radio podcast, researcher Dipto Barman, who recently completed his PhD at Trinity College Dublin, sheds light on the rise of deep fakes and disinformation, especially during sensitive periods like elections. With heightened tensions and underlying agendas, Barman emphasises the need for caution when consuming online content, particularly in relation to major global events.
Barman explains the technology behind deep fakes, which are synthetic media created using AI and machine learning, primarily through Generative Adversarial Networks (GANs). He highlights the dangers of deep fakes in manipulating public opinion, citing recent disinformation cases, such as false videos circulating in the context of political conflicts, like the Israel war.
However, he also discusses the ways in which AI can be used to counteract deep fakes, including AI-based detection algorithms that identify inconsistencies, such as unnatural facial movements or shadow errors. Blockchain technology is also being leveraged to verify the authenticity of digital content by tagging outputs with metadata. Additionally, digital forensic tools are employed to analyse data compression at the pixel level, identifying discrepancies that reveal manipulated content.
Tune in to the ADAPT Radio podcast for insights into how we can increase media literacy and use technology to combat the spread of disinformation. Listen here or wherever you get your podcasts.
