SiiP

Programme: FP7 - SEC

Instrument: Collaborative Project

Total Project budget: € 15.126.660,20

Duration: 48 months (May 2014- April 2018)

Full project title: SiiP: Speaker Identification Integrated Project

---

RUG responsibilities:

In the SiiP Project RUG is Work Package leader of:

WP2 Legal, Ethics And Societal Aspects. The objectives of this WP are:

To identify, analyse, and assess ethical, legal and social issues arising in the use of speech recognition technology in the field of criminal justice and to distinguish between ethical, legal and social issues arising at three distinct stages: development of technology; application of technology; sharing of data.

To interact continuously with technology developers and end-users both in order to contribute ethical, legal and social analysis to the work of technology workpackages and in order to develop, disseminate, and gain feedback on ethical, legal and social research.

To produce practical guidance for end-users of speech identification technologies on ethical, legal and social risks of using such technologies in their work.

To provide the R&D work throughout with the appropriate legal framework, ensuring end-products are fully compatible/ compliant with current trends in European and international privacy and data protection standards implementing the internationally-endorsed foundational principles of Privacy by Design.

Project Summary:

SIIP is a break-through Suspect Identification solution based on a novel Speaker-Identification (SID) engine fusing multiple speech analytic algorithms (e.g. voiceprints recognition, Gender/Age/Language/Accent ID, Keyword/ Taxonomy spotting and Voice cloning detection). This Fused Speaker Identification will result in significantly higher true-positive speaker identification, reduced False-Positives/Negatives while increasing reliability & confidence. SIIP analyzes rich metadata from voice samples and social media. SIIP provides judicial admissible evidence for identifying crime/terror suspects as well as for mapping/tracing the suspect terror/crime network. SIIP is crucial when individuals use Internet-based applications (e.g. VoIP or social media) to plan a crime or terrorist attack. SIIP's results can easily be shared with relevant authorities based on a sustainable SIIP Info Sharing Center (SISC) located at INTERPOL. SISC guarantees an increased reliability of the identification results through advanced technologies and through voice samples checked against a large centralized database of samples collected by INTERPOLs 190 members (based on standard operating/data privacy procedures). SIIP multiplies and increases the information sharing and cooperation in the LEA community and speeds up the use of Speaker Identification by LEAs in Europe not only for individual identification but also for authentication. SIIP runs on all speech sources (e.g. Internet, PSTN, Cellular and SATCOM) and uses the latest OSINT data mining applications to obtain and corroborate voice samples. The SIIP consortium consists of 17 partners bringing together end-users, SME’s, industrial and academic partners from a variety of fields including Speech analytics, Social Media Analytics and Integration. To maximize its impact, SIIP will be designed, developed and tested with INTERPOL and police forces in the UK and Portugal, taking into account the various EU legal/ethical aspects and Interpol regulations.

For further information please contact: Melania Tudorica.