Here are some characteristics of good and bad assessment in the age of artificial intelligence. These characteristics emerged from a preliminary analysis of university students’ comments as part of the PUPP data collection.

Characteristics of a good activity	Characteristics of a bad activity
Contains changes in each session	Has already been used during from a previous session
Contains complex and grounded issues​​	Contains only factual questions or simple exercises repetitive
Need a creativity on the part of the student	Contains predictable questions
Encourages the application of knowledge	Focuses on memorization
Promotes creation​	Lack of context
Promotes student engagement​	Does not arouse interest in the student
Can with difficulty be answered by AI	Can easily be answered by answers generated by AI
Combine several shapes assessment ( oral , written , individual , group )	Uses a single assessment format
Goals clear , precise which are often repeated or presented in various ways	Poorly defined objectives , lack of specificity
Integrates elements​ evaluation in a way frequent and continuous	Assessment focused towards the final result
Offers choices​ staff to students to make their more meaningful learning	Lack of personalization for learners or lack of contextualization
Student Collaboration framed with a distribution clear responsibilities​	None instructions on the division of tasks in teamwork
Assessment supervised during high – stakes assessments	Lack of oversight during high- stakes assessments
Contains a verification of authenticity	None measure required to verify the authenticity of the work
Contains a reflection metacognitive on the process of creating work	Lack of reflection on the process of creating work
Contains AI training elements​	Lack of training on the use of AI