Australians Struggle to Identify AI-Generated Deepfakes, Commbank Survey Finds

Over half of all Australians are unable to distinguish between a real image and an AI-generated deepfake, according to new data.\n\nA Commbank survey of nearly 2000 individuals across Australia found that 89% of respondents expressed confidence in their ability to identify an AI-generated scam. However, only 42% were actually successful in differentiating between genuine and AI-generated images.\n\nJames Roberts, Commbank's General Manager for Scam and Fraud, stated that these findings reveal a growing disparity between public confidence and reality. He noted that scammers are increasingly utilizing AI to target both everyday Australians and small businesses.\n\nScammers are employing deepfake photos, voice and text clones of loved ones, senior executives, and government officials, as well as fabricated investment videos, to execute their schemes.\n\nProfessor Monica Whitty, a Professor of Human Factors in Cyber Security at Monash University, explained that these types of scams exploit human tendencies to trust faces, voices, and familiar people. Commbank's research identified investment scams as the most common, followed by business email compromise scams and relationship scams.\n\nThe Australian Competition & Consumer Commission's (ACCC) National Anti-Scam Centre reported that Australians lost nearly $260 million to scams between January and September of the previous year. Online content, including fake websites, advertisements, social media, and applications, was cited as the most frequent method of scam delivery.\n\nThe National Anti-Scam Centre advises Australians to pause and verify information before making financial decisions, recognize warning signs, communicate with loved ones about scams, and engage in open discussions regarding potential threats. They caution against rushing decisions about money or sharing personal details, especially when a sense of urgency is imposed.\n\nRoberts affirmed that the established method of stopping and checking for scam indicators remains the most effective defense against AI-powered scams. He suggested measures such as asking for a family safe word or contacting loved ones directly via their usual phone number when encountering suspicious messages, like