Tremendous challenges remain in terms of data quality, repeatability and accuracy, and related high costs in pavement data collection and processing. Out of the four major pavement evaluation categories (function, condition, structure, and safety), this proposal focuses on developing innovative Artificial intelligence (AI) solutions to condition and safety evaluation. The objective is to provide a generational advancement of sensors and solutions for the rapid and accurate evaluation of multiple pavement distresses and safety properties based on Deep­ Learning (DL) networks and sub-mm 3D laser imaging sensors. Cracking is one of many pavement distresses. Based on the proposal team’s prior work on DL based method (CrackNet) for automated cracking surveys, a universal new DL network is proposed to conduct multiple­ distress surveys in a single pass. In addition, based on the preliminary research by the team on using a non-contact 0.1-mm resolution 3D laser imaging sensor (Safety Sensor), the proposal will develop DL based Super-Resolution techniques to reconstruct pavement surfaces at 0.1-mm resolution in 3D so that highway speed survey of pavement micro- and macro-texture, and friction can be a reality. The new AI methods have the potential to retire the traditional water-­based and contact-based friction testers, and define new standards in safety surveys.