Deepfake statistics 2026: key facts for Australian CFOs and finance leaders

A deepfake is synthetic media created with artificial intelligence, typically video, audio, or images altered to show someone saying or doing something that never happened. Fraudsters use the same technology to impersonate executives, vendors, and colleagues in payment scams. Read more on how deepfake AI works.

Estimates put online deepfake volume at roughly 500,000 in 2023, rising to approximately 8 million by 2025, a figure also cited by the UK government. The pace of growth means any fixed count is out of date within months.

Deepfake fraud is the use of AI-generated audio or video to impersonate a real person, usually an executive, vendor contact, or colleague, in order to authorise a fraudulent payment or extract sensitive information. Read the full breakdown of how deepfake fraud works.

Creating and sharing non-consensual sexually explicit deepfakes is a specific criminal offence in Australia, covered by Commonwealth Criminal Code amendments together with separate state laws in jurisdictions including New South Wales and South Australia, with penalties varying by jurisdiction. Deepfake-enabled fraud, such as executive impersonation used to authorise a payment, isn't a separately named offence. It's prosecuted under existing fraud and deception laws instead, which apply regardless of the method used to deceive. For finance teams, this distinction matters in practice: prosecution can follow after a loss occurs, but it doesn't recover a payment already sent to a fraudster's account, which is why independent verification before payment remains the more reliable safeguard.