A voice identification system used by the Australian government has been found to have a serious security flaw. The system, known as voiceprint, allows individuals to verify their identity over the phone by using their voice along with other information. However, an investigation by Guardian Australia has revealed that the system can be fooled by an AI-generated voice. Using just four minutes of audio, a journalist was able to generate a clone of their own voice and gain access to their own Centrelink self-service account. This vulnerability highlights the limitations of biometric identification as a security tool.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| Voice identification system has a security flaw | From vulnerable voiceprint system to secure identification methods | Improved voiceprint technology and stronger security measures | Advancements in AI and biometric technology |
| AI-generated voice can fool the voiceprint system | AI-generated voices can bypass biometric identification | Enhanced detection of synthetic voices and improved security measures | Advancements in AI and voice cloning technology |
| Journalist clones their own voice to access their own account | Increased vulnerability of personal information | Stricter verification methods and improved security measures | Ease of access to voice cloning technology |
| Voiceprint system used by millions of Centrelink clients | Secure authentication method using voiceprint | Enhanced security measures and improved authentication methods | Need for convenient and secure identification |
| Biometrics are not foolproof for security | Limitations of biometrics as a security tool | Stricter legislation and standards for biometric verification | Balancing convenience and security |
| Government agencies need better biometric systems and legislation | Improved systems and legislation for biometric verification | Stricter legislation and basic standards for safe and reliable technology | Ensuring safety, reliability, and protection from cybercrime |
| Companies claim robust measures to protect against AI voice cloning | Protection against AI voice cloning threats | Continuous monitoring, system improvements, and staying ahead of new threats | Advancements in AI and machine learning |
| Voice cloning technology is becoming more accessible | Increased vulnerability to voice reproduction | Awareness of public voice recordings and potential risks | Accessibility and affordability of voice cloning technology |