printer friendly version

IoT and behavioural authentication

IoT represents an increasing security risk to compromise your most personal security credentials.

Security begins at home and it begins with you, the individual. IoT represents an increasing security risk to individuals in the form of pervasive, always-on monitoring of your personal activity with a potential compromise of your most personal security credentials.

Consider the four factors of authentication. First and foremost is something you know – a password for example. The second is something you have, such as an access card, key fob or hardware-based token generator. Third is something you are, such as your iris pattern or fingerprint – both are typical biometrics. The fourth factor of authentication is something you do – the way you walk, breathe, speak or any other manner of activity which has a recognisable and unique pattern over time. This latter is called behavioural authentication and while this factor is often tied into general biometrics, it is distinctly different.

It’s not unimaginable at this point that your fourth factor of authentication can be reconstructed from data gathered by IoT devices. Think about how Internet-connected shoes can learn intimate details of their wearer’s gait; or the smartwatch that learns the specific timing of its wearer’s heartbeat. Then there is the speaker-based home assistant that learns the pitch and timbre of the home occupants’ voices.

Post quantum security

The first two factors of authentication rely on a number, key or phrase that is susceptible to brute force attacks to crack the credential. While still a long way off, the immense power of quantum computing will undermine the security of the first two factors of authentication. The advent of quantum computing will mean that technology manufacturers will come to increasingly rely on the third and fourth authentication factors.

When the encryption mechanisms behind passwords and OTPs become insecure, you will certainly find yourself logging into your Internet banking with a passphrase spoken into your microphone. Your typing style, based on flight, sequence and pressure, will become your ongoing signature. These voice and keyboard behaviours are just two patterns that are in use today and will become increasingly important.

Just like you would not be careless with your house keys, it’s important to know the value your habits will one day have. In a situation where the data collected by IoT devices compromises users’ habits, those people would lose the ability to secure their devices and their logins.

For example the first known financial scam using synthesised voice was reported in September 2019[1] when scammers replicated a CEO’s voice to authorise payments totalling R3,6 million from a company. This was done using deep fake (a technique for human image synthesis based on artificial intelligence)-derived AI and various voice recordings of the victim.

Commitment to privacy

There have been some positive developments in the privacy space, following increased scrutiny from legislators in the form of GDPR/PoPIA and general public awareness around exploitation of personal data. The various social media data abuse scandals over the years are also turning public sentiment. People are becoming more aware and protective of their personal data which will extend to behavioural data as well.

Manufacturers of IoT devices are also showing encouraging signs, for example Amazon has added auto delete features to its Alexa range after previously admitting that audio recordings are stored indefinitely[2]. Moreover, Apple changed its policy to no longer retain audio recordings by default. It’s worth taking into serious consideration the fact that Google’s chief of devices and services recommended that one should disclose if one’s home has smart-home devices installed before issuing invitations to guests[3].

As IoT devices proliferate, vendors must commit to transparency about what data they collect and how they use it. A firm legislative effort to curb needless collection of data, especially around sensitive activity and habit patterns, will go a long way to securing the market for a post-password authentication world. Until then, buyers of IoT devices need to read the fine print and make sure they know what behavioural data is being collected and how it is used.

Gregory Dellas

[1] https://threatpost.com/deep-fake-of-ceos-voice-swindles-company-out-of-243k/147982/

[2] https://www.theverge.com/2019/7/3/20681423/amazon-alexa-echo-chris-coons-data-transcripts-recording-privacy

[3] https://www.bbc.com/news/technology-50048144

Share this article:

Further reading: