An allophone is a short speech-sound unit used to build synthetic speech from fragments. Instead of storing whole words or sentences, a speech synthesiser can combine these small pieces into audible output. This approach was common in early voice hardware because it saved memory and processing power.
In cyber security, allophones matter because they show how constrained devices generate voice. Legacy embedded systems, kiosks, and IVR platforms may still rely on fixed speech units or similarly rigid synthesis pipelines, which can affect quality, language coverage, and spoofing resistance. Attackers can abuse synthetic speech in social engineering and voice phishing, while defenders may inspect the speech stack to understand how prompts are generated, whether audio can be manipulated, and where a device’s limited design creates security or reliability gaps.