Patent attributes
An authentication subsystem receives a request for authentication of a user to access a server. A voiceprint is determined for the user based on recordings of the user speaking. Authentication phrases is determined for the user based on the voiceprint. Based public user data and/or user transaction data, user knowledge is determined which reflects topics about which the user has knowledge. An authentication question is generated based on the authentication phrases and the user knowledge. An anticipated answer to the authentication question one of the authentication phrases. The authentication question is provided to a device operated by the user. If a response voiceprint associated with an actual spoken response to the authentication question sufficiently matches an anticipated voiceprint for an expected recording of the anticipated answer, the user is granted access to the server. Otherwise, access to the server is denied.
