Increasing levels of communication and automation has led to the development of cyber physical systems applications known as cooperative driving automation. While such applications provide enormous benefits in improving transportation systems operation and environmental sustainability, they also introduce new risks to smart cities. These systems provide wider access to adversaries and increase the attack surface to compromise the operation of smart cities. This study develops a vehicle to infrastructure (V2I) based cooperative driving automation platform to understand and analyze the cyber risks in these systems from a V2I perspective in a representative traffic environment. Three cyberattacks (Fake BSMs, replay and denial of service) were emulated as a cooperative game to assess the impact on volatility, energy efficiency and traffic flow stability. The critical nature of cooperative driving automation is revealed by this research and flow stability, and energy efficiency of cooperative driving automation are severely impacted by the three categories of cyberattacks. Further, a 35 %-40 % increase in volatility in acceleration-deceleration regimes along with 20 % increase in emission is observed as opposed to the baseline case with no cyberattacks. The proposed long-short term memory neural networks (LSTM) showed an average detection accuracy of around 98 % for the analyzed cyberattacks. The research findings would foster the development of algorithms to continuously monitor the state of cooperative automated driving environment for anomalous behavior and promote resilient and environmentally sustainable cities in the future.