In human communication, speech is a medium to impart and exchange information. Speech encodes a wealth of information on, e.g.: socio-economic background; racial, ethnic, and geographical origin; religious, political, and philosophical beliefs, health and emotion condition; identity; gender, and age. All of which is recognisable by humans. With speech technology evolving, machines are increasingly able to perceive and process this information. In certain usage scenarios, this specific information is needed to guarantee functionality and a proper service. But in many usage scenarios, this information is not needed and should be neglected.
For instance, detecting early signs of Parkinson’s disease, depression/suicide risk, and dementia help physicians sustain the health of a person by comprehensive medical profiling. In other circumstances, such profiling, while possible, is regarded as morally wrong indicating the importance of privacy in the context of speech analysis. Therefore, the terms Personally Identifiable Information (PII) and Public Sector Information (PSI) are used. [Through PII as information representation, the identity of an individual can be reasonably inferred directly or indirectly. Information produced by public entities is PSI.] Regarding privacy in speech communication, we need not only to arrive at a better understanding of when speech data is seen as PII or PSI but also to which extent PII is needed for an application - and when we want to refrain from it.
We, furthermore, need to secure epistemic value across disciplines to arrive at a mannerism that is credible and adequate when designing speech & language technology (SLT) that respects fundamental human rights throughout product and service development, implementation, and integration. Moreover, the interoperability of privacy and security needs to be better understood. Security is used to countermeasure subversive users, such as detecting fake audio data in online banking using voice biometrics and call-center fraud prevention. Privacy is used to protect individuals from harm, e.g., through unbeknownst linkage of their (speech) data resulting in subsequent harm when harvested by adversaries. In the age of virtual voice assistants, speech communication is intimately related to daily life in digital societies, and so are security and privacy matters as well. Consequently, we need to also elaborate on how to involve the users and their perspectives on questions of privacy and security.
This special issue seeks to provide a venue for ongoing research in the recently formed research community on Security & Privacy in Speech Communication (SPSC), where views of technological and humanities communities nurture one another to develop multidisciplinary and interdisciplinary skills.
Technology and research need to aid individuals and society for which scientific inquiry needs to channel durable solutions. Currently, academic disciplines exist in silos; fields need to be bridged, and real-world problems to be solved. Ways to achieve this can be many in the anticipation of currently lacking multidisciplinary and interdisciplinary skill sets, e.g.: concise summaries of developments within a field for their relation to SLTs (hypothetical or existing); portrayals of community-driven efforts fostering security and/or privacy in SLTs; insights into the users’ ideas of the concepts and its effects on their usage; syntheses and analyses of SLT systems and continuous holistic improvement of their design, and security & privacy solutions for SLT applications including user experience(s).
The topics of interest for the special issue include, but are not limited to:
- Sustainability and Economics
- Secure/Private Speech Communication
- Secure Computation and Cryptography
- Psychology, e.g. media psychology and social psychology
- Privacy-preserving Human-Computer Interfaces
- Policy and Governance, Law, and Ethics
- Natural Language Processing
- Machine Learning for secure and private speech processing
- Law Enforcement and Forensics
- Information and Communication Technology for secure and private speech processing
- Human-Computer Interaction Perspective on Security and Privacy
- Digital Humanities and Anthropology regarding privacy and security needs
Submission deadline: 30 May 2023
Lead Guest Editor
Rodrigo Capobianco Guido, São Paulo State University, Brazil,
Astrid Carolus, Julius-Maximilians University Würzburg, Germany
Hemant A. Patil, DA-IICT, India
Ingo Siegert, Otto-von-Guericke Universität Magdeburg, Germany
Important: Authors should select "Security & Privacy in Speech Communication" when they reach the “Article Type” step in the submission system.
Submission instruction: The submission system will open in early January 2022. Submissions deemed suitable will be reviewed by experts in different disciplines. Published articles will appear in different regular issues of the journal (clearly marked and branded as Special Issue articles. Please use IEEE style for citations; footnotes are neat for brief clarifications, and the layout should support reading (double column is a suggestion only). Write what prevails, make it digestible, and palate it to your audience of inquiry (impart knowledge, empower understanding, and take care of the willing reader to mature).
Submission guidelines: https://asmp-eurasipjournals.springeropen.com/submission-guidelines