Intelligent collaborative robots or smart cobots can achieve high levels of flexibility by combining the human ability to adapt to new tasks with the performance of automated robots (precision, repeatability, etc.). This is a major innovation for Industry 4.0. Nevertheless, at present, cobots are not widely deployed in industry because they are difficult to evaluate and current standards for them are limited. The evaluation of collaborative tasks is difficult due to their specificities such as the variability of human behavior, artificial intelligence systems (e.g., smart cameras), and advanced control laws. This paper is a short review that aims to identify the methods and material resources needed to address this need for evaluation of intelligent cobots to foster their development and acceptability.