Anthropic’s Project Glasswing, a program evaluating the security implications of advanced AI models, includes Australia and highlights how quickly these capabilities are evolving. Frontier AI systems from companies like Anthropic, OpenAI, and Google allow cyber criminals to generate highly personalised communications at scale, replicate writing styles, and create authentic-looking messages, making traditional scam red flags disappear. Eftsure research found that 82 per cent of Australian finance professionals are concerned about AI-assisted scams in the future, while 72 per cent believe they are more at risk of fraud today than they were a year ago. The article notes that AI also offers the ability to continuously monitor transactions, identify anomalies, validate payment information, and surface risks, with organisations gaining the most using AI to bring greater certainty to high-risk financial decisions before money leaves the business.