Voice-based Decentralized Identity

Blockchain technology makes possible a secure and verifiable implementation of decentralized identity - an identity that empowers users by providing them with control of their own identity data. Unleashing the power of decentralized identities requires work on interfaces that users will use to create and manage their identity and associated data. Based on the Global ID4D Dataset, 1 billion people lack official proof of identity. Most of these people are illiterate and often lack smartphones, computers, and broadband internet access. But, most of them have access to a 2G network and are capable of using the services. Our solution leverages 2G network voice call functionality to give these people a suitable interface for identity enrollment, authentication, and management. By leveraging voice biometrics and secure enclaves, we make it possible to generate and protect a unique key pair, which is a starting point for building a secure and verifiable decentralized identity.

The solution is offering a globally viable alternative to digital identity enrollment, authentication, and management interfaces that by design preclude broad-base participation ,especially  for vulnerable groups who often lack smartphones, computers, broadband internet access, and a sufficient level of literacy. Based on the Global ID4D Dataset, 1 billion people lack an official proof of identity, 81% of these people live in Sub-Saharan Africa and South Asia while over 75% of the world's 781 million illiterate adults are found in South Asia, West Asia, and sub-Saharan Africa. Therefore, an interface that would potentially be inclusive for all populations should be usable without an expensive and complex device (smartphone, computer…), broadband internet access, or a material  level of literacy. In Sub-Saharan Africa for example, more than 90% of the population has, at least, access to a 2G network and are capable of using voice call capability regardless of their level of literacy. So, using the voice call as an interface to digital identity will allow the digital inclusion of most of the vulnerable groups. By leveraging voice biometrics and secure enclave, our solution uses voice calls as an entry point to the digital while ensuring security, trust, robust privacy, and data protection.

The solution serves all people that have an interest in getting and managing a digital identity but are blocked by a lack of literacy, broadband internet access, and expensive and complex devices like smartphones and computers. As long as they have access to a 2G network and are capable of making a voice call, they will have the capabilities of creating in a privacy-preserving way a decentralized identity.

A decentralized identity puts the user at the center of the ecosystem with the ability to control and maintain their own identity. It is inclusive, secure, useful, fit for purpose and moreover it gives to the user the choice.

Every transaction starts with an identity. A verifiable and trusted self-managed digital identity will allow these people access to new services that will improve their lives.

Our solution tackles both challenge dimensions in the Global Prize. Our voice-based identity interface allows people even in vulnerable groups to register for a secure and verifiable ID in a privacy-preserving way that helps protect their health. It also offers two levels of authentication using a voice call. The first level is based on biometric verification using the input voice and the user’s voiceprint registered during the enrollment process. The second level is based on verification from a blockchain of credentials associated with the user’s unique identifier and received from trusted entities in the network like a government identity agent.

Today, most decentralized identity implementations use digital wallets (mobile applications) as user interfaces. These implementations require a smartphone and internet access to work properly. These dependencies create a limitation on their use by vulnerable groups who often lack smartphones, computers, broadband internet access.

https://www.youtube.com/watch?v=bkvxYkeh71E

Our solution leverages voice biometrics to offer an interface that gives people the possibility to create and manage their identity using voice calls. Besides bringing uniqueness, voice biometrics allows users to manage their identity from any phone, regardless of phone number, and from anywhere with a 2G network coverage. The voice verification technology has evolved significantly in these past years with solutions that can be used in any language or dialect and with presentation attack detection methods, capable of distinguishing between a live voice and a recorded, synthetic or computer-generated version of the voice. Our main innovation is how our identity cloud agent that receives the call, uses fuzzy extractors to generate from the voice biometric input a unique private cryptographic key that is the signing key associated with the user identity. The key generation is done in a secure enclave that helps protect user sensitive data with respect to confidentiality and integrity.

Our solution uses voice biometrics, fuzzy extractors, and secure enclaves to offer an inclusive, secure and privacy-preserving interface for identity registration and management.

Voice Biometrics uses physiological and behavioral aspects of the voice to identify users. It is readily adopted because people are familiar with it, and it’s easy to use. It also doesn’t raise concerns about hygiene or cultural appropriateness, since the voice capture process requires no physical contact between individuals and a device, and no physical intervention by operators. Setting up a voice-based ID system is cost-effective because it requires little capture hardware beyond a simple microphone for recording voices, and associated matching technology is on par with other modalities. The 2019 NIST Speaker Recognition Evaluation CTS Challenge indicates great progress in speaker recognition technology, with relative performance improvements as large as 70% for the leading system.

https://www.nist.gov/publications/2019-nist-speaker-recognition-evaluation-cts-challenge

Fuzzy extractors provide key generation from biometrics and other noisy data. They extract a uniform and random string ? from an input ? with a tolerance for noise. If the input changes ?’ but is still close to ?, the same string ? will be re-constructed.

https://ro.uow.edu.au/cgi/viewcontent.cgi?referer=https://www.google.com/&httpsredir=1&article=1698&context=eispapers1

https://www.cs.bu.edu/~reyzin/papers/fuzzy.pdf

A trusted execution environment (TEE) or Secure Enclave is a secure area of a main processor. It guarantees code and data loaded inside to be protected with respect to confidentiality and integrity. A Secure Enclave as an isolated execution environment provides security features such as isolated execution, integrity of applications executing with the TEE, along with confidentiality of their assets.

https://www.intel.fr/content/www/fr/fr/architecture-and-technology/software-guard-extensions.html