Artificial intelligence is rapidly reshaping how students learn, how knowledge is constructed and how universities assess understanding and skills. For higher education, the challenge is not simply technological adoption but ensuring that AI strengthens learning, protects academic standards and prepares graduates for an increasingly AI-enabled world. While students can readily identify their best learning experiences, often citing inspiring teachers or practical activities, they frequently struggle to identify their best assessment experience. This gap raises questions about how assessment could be designed and experienced well in the AI-era.

This paper reports on a series of interactive workshops held during The Open University's Student Consultation Week, where students were invited to ‘design their dream assessment’ in the AI era. To prepare students, they were introduced to Black and Wiliam’s (2012) model of assessment for learning and The Open University’s Academic Integrity Principles for Assessment Design. They were also shown recent student AI usage data (Stephenson & Armstrong, 2026) and asked to reflect on their own AI-supported learning practices. Working in small groups, students selected a topic and level, articulated a learning outcome, and designed an assessment that incorporated AI capability building while enabling markers to see evidence of the learning process and responsible AI use.

Student designs give institutions insight into what motivates students to genuinely engage, how they already use AI for assessment, where their human-AI boundaries are and what they consider is fair.