Callisto: A Cryptographic Approach To Detect Serial Predators Of Sexual Misconduct

Abstract

Callisto, a non-profit that has created an online sexual assault reporting platform for college campuses, has expanded its work to combat sexual assault and professional sexual coercion in other industries. In our new product, users will be invited to an online matching escrow that will detect repeat perpetrators and create pathways to support for victims. Users of this product enter incident details and perpetrator iden- tities into the escrow. This data can only be decrypted by a Legal Options Counselor (a third-party lawyer vetted by Callisto) when at least one other user enters the iden- tity of the same perpetrator. If perpetrator identities match, each user is assigned a Legal Options Counselor, who will connect users to each other (if appropriate) and help each user determine their best path towards justice. User relationships with Le- gal Options Counselors are structured so that relevant communications benefit from client-counselor privilege. A combination of client-side encryption, encrypted com- munication channels, oblivious pseudo-random functions, key federation, and Shamir Secret Sharing keep data encrypted so that only Legal Options Counselors gain access to identifying user submitted data when a perpetrator match is identified. In this paper, we present an informal risk management assessment, threat model, and cryp- tographic solution overview for our new product. A later paper will provide a formal security analysis and mathematical proofs of our cryptographic scheme.

(Taken from the paper cited below)

References