Security

In May 2018, researchers from the University of California, Berkeley, published a paper that showed audio commands undetectable for the human ear could be directly embedded into music or spoken text, thereby manipulating virtual assistants into performing certain actions without the user taking note of it. The researchers made small changes to audio files, which cancelled out the sound patterns that speech recognition systems are meant to detect. These were replaced with sounds that would be interpreted differently by the system and command it to dial phone numbers, open websites or even transfer money. The possibility of this has been known since 2016, and affects devices from Apple, Amazon and Google.

In addition to unintentional actions and voice recording, another security and privacy risk associated with intelligent virtual assistants is malicious voice commands: An attacker who impersonates a user and issues malicious voice commands to, for example, unlock a smart door to gain unauthorized entry to a home or garage or order items online without the user's knowledge. Although some IVAs provide a voice-training feature to prevent such impersonation, it can be difficult for the system to distinguish between similar voices. Thus, a malicious person who is able to access an IVA-enabled device might be able to fool the system into thinking that they are the real owner and carry out criminal or mischievous acts.