Keynotes: WDC

Temporal Evolution of Human Perceptions and Detection of Deepfakes: An Empirical Study



Prof. Ganna Pogrebna
Executive Director/Professor (Honorary)
The Artificial Intelligence and Cyber Futures Institute, Charles Sturt University and The University of Sydney

Abstract: This study provides an extensive review of how human perceptions of deepfakes have evolved from 2018 up to 2022. Using a variety of qualitative and quantitative methodologies, we explore shifts in public understanding, awareness, and ability to identify deepfakes over this period. We also investigate the influence of the ‘Big Five’ personality traits on the ability to detect deepfakes, offering novel insights into the psychological factors underpinning individual variability in detection success. Our findings suggest a significant increase in public awareness of deepfakes over the study period, although this has not been accompanied by a parallel increase in detection ability. Despite the evolution of deepfake technology, many individuals still struggle to differentiate between real and manipulated media, with false negatives and positives commonly reported. Furthermore, our results highlight the impact of individual personality traits on deepfake detection. Openness and conscientiousness are associated with higher detection success, potentially due to increased curiosity about emerging technologies and a meticulous approach to detail, respectively. These findings underscore the urgent need for robust public education strategies and the development of advanced detection tools to combat the growing influence and sophistication of deepfakes. Moreover, our study provides a foundation for future research examining the complex interplay between personality traits and susceptibility to digital misinformation.
Speaker’s Bio: Ganna Pogrebna is a pioneer in behavioural data science. She is Executive Director of the Artificial Intelligence and Cyber Futures Institute at Charles Sturt University, Honorary Professor at the University of Sydney, and Lead for Behavioural Data Science at the Alan Turing Institute. Blending behavioural science, AI, computer science, data analytics, engineering, and business model innovation, Ganna helps businesses, cities, charities, and individuals to better understand why they make decisions they make and how they can optimize their behaviour to achieve higher profit, better social outcomes, as well as flourish and bolster their wellbeing. Her recent projects focus on smart technological and social systems, cybersecurity, human-computer and human-data interactions, and business models. Her most impactful projects concentrated on cybersecurity as a behavioural science as well as applications of behavioural data science to media industry . Her digital security risk-tolerance scale (CyberDoSpeRT) is widely used in Australia and abroad. Ganna’s contributions to risk analytics and modelling was recognized by the Leverhulme Research Fellowship award. In January 2020, she was also named as the winner of TechWomen100 – the prize awarded to leading female experts in Science, Technology, Engineering and Mathematics in the UK. She is also named as one of 20+ Inspiring Data Scientists by the AI Time Journal. Ganna runs the Data Driven blog on YouTube as well as Inclusion AI blog. Her work is regularly covered by the traditional as well as social media. Ganna is one of the contributors to the Oxford Handbook of AI Ethics. She is also currently co-editing the Cambridge Handbook of Behavioural Data Science, which is due to be published in 2023 by Cambridge University Press.