Attack and Defence Modelling for Attacks via the Speech Interface

Bispham, Mary; Agrafiotis, Ioannis; Goldsmith, Michael

doi:10.5220/0007469705190527

Cited by 6 publications

(12 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This section presents experimental work showing that it is possible to hide malicious voice commands to the voice-controlled digital assistant Google Assistant in word sounds that are perceived as meaningless by humans. We term this type of attack a 'nonsense' attack, in accordance with a taxonomy published in a previous paper [4], which categorises attacks via the speech interface according to human perceptual categories. The attack can also be characterised as a black-box adversarial learning attack.…”

Section: Description and Contextmentioning

confidence: 99%

“…The space of meaningful word sounds within a language at a given point in time is generally confirmed by the inclusion of words in an established reference work, such as, in the case of English, the Oxford English Dictionary. 4 Word sounds that are outside this space can be described as nonsense words. Nonsense words are a grey area between non-speech, i.e.…”

Section: Description and Contextmentioning

confidence: 99%

“…This section presents experimental work demonstrating that it is possible to gain unauthorised access to a voice-controlled system using utterances that are accepted by the system as a target command despite having a different meaning to the target command in terms of human understanding. We term this type of attack a 'missense' attack, in accordance with a taxonomy of attacks via the speech interface published in a previous paper [4], which categorises attacks via the speech interface according to human perceptual categories. The attack can also be characterised as a black-box adversarial learning attack.…”

Section: Description and Contextmentioning

confidence: 99%

See 2 more Smart Citations

Black-Box Attacks via the Speech Interface Using Linguistically Crafted Input

Bispham

Rensburg

Agrafiotis

et al. 2020

Communications in Computer and Information Science

Self Cite

View full text Add to dashboard Cite

This paper presents the results of experiments demonstrating novel black-box attacks via the speech interface. We demonstrate two types of attack that use linguistically crafted adversarial input to target vulnerabilities in the handling of speech input by a speech interface. The first attack demonstrates the use of nonsensical word sounds to gain covert access to voice-controlled systems. This attack exploits vulnerabilities at the speech recognition stage of handling of speech input. The second attack demonstrates the use of crafted utterances that are misinterpreted by a target system as a valid voice command. This attack exploits vulnerabilities at the natural language understanding stage of handling of speech input.

show abstract

Section: Description and Contextmentioning

confidence: 99%

Section: Description and Contextmentioning

confidence: 99%

Section: Description and Contextmentioning

confidence: 99%

See 1 more Smart Citation

Black-Box Attacks via the Speech Interface Using Linguistically Crafted Input

Bispham

Rensburg

Agrafiotis

et al. 2020

Communications in Computer and Information Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…Bispham et al [10] have developed a taxonomy of potential attacks via the speech interface that is organised according to the nature of the attack in terms of human perception. The taxonomy divides such attacks into two high-level categories: overt attacks, which aim to take control of a target system using plainspeech voice commands; and covert attacks, in which malicious voice commands are concealed in a cover medium so as to make them imperceptible to human listeners.…”

Section: Prior Work On the Security Of The Speech Interfacementioning

confidence: 99%

“…This paper is an extended version of a previous paper in which our framework for modelling the security of the speech interface was first presented (see Bispham et al [7]). The work presented here extends the work in the earlier paper with the proposals for the development of new defence mechanisms referred to above.…”

Section: Introductionmentioning

confidence: 99%

The Security of the Speech Interface: A Modelling Framework and Proposals for New Defence Mechanisms

Bispham

Agrafiotis

Goldsmith

2020

Communications in Computer and Information Science

Self Cite

View full text Add to dashboard Cite

This paper presents an attack and defence modelling framework for conceptualising the security of the speech interface. The modelling framework is based on the Observe-Orient-Decide-Act (OODA) loop model, which has been used to analyse adversarial interactions in a number of other areas. We map the different types of attacks that may be executed via the speech interface to the modelling framework, and present a critical analysis of the currently available defences for countering such attacks, with reference to the modelling framework. The paper then presents proposals for the development of new defence mechanisms that are grounded in the critical analysis of current defences. These proposals envisage a defence capability that would enable voice-controlled systems to detect potential attacks as part of their dialogue management functionality. In accordance with this high-level defence concept, the paper presents two specific proposals for defence mechanisms to be implemented as part of dialogue management functionality to counter attacks that exploit unintended functionality in speech recognition functionality and natural language understanding functionality. These defence mechanisms are based on the novel application of two existing technologies for security purposes. The specific proposals include the results of two feasibility tests that investigate the effectiveness of the proposed mechanisms in defending against the relevant type of attack.

show abstract

Compromised Computers Meet Voice Assistants: Stealthily Exfiltrating Data as Voice over Telephony

Rajput

Ahamad

2021

2021 51st Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN)

View full text Add to dashboard Cite

Attack and Defence Modelling for Attacks via the Speech Interface

Cited by 6 publications

References 0 publications

Black-Box Attacks via the Speech Interface Using Linguistically Crafted Input

Black-Box Attacks via the Speech Interface Using Linguistically Crafted Input

The Security of the Speech Interface: A Modelling Framework and Proposals for New Defence Mechanisms

Compromised Computers Meet Voice Assistants: Stealthily Exfiltrating Data as Voice over Telephony

Contact Info

Product

Resources

About