Rethinking assessment

Rethinking assessment:

Teachers imagine new frameworks for success in a GenAI world

Has generative artificial intelligence (GenAI) made the classroom assignment obsolete? At the Eureka! Institute, we wanted to lean into the discomfort posed by this question, taking seriously the need to probe our existing methods of assessment as we reckon with the opportunities and threats posed by GenAI. We propose that the current moment offers a unique opportunity for educators to re-think not only how they assess, but what should be valued in assessment. 

To that end, we arranged for a group of past and current teachers at UTS, working across subject areas, to reflect on how they see their assessment practices and educational philosophies changing in light of GenAI. Some of the questions we asked were: What should we evolve so that assessment in the classroom continues to be meaningful? How should we re-imagine ‘success’ as generative AI redefines what it means to achieve? What can teachers let go of in this new era of assessment?

Melissa Shaddick, a drama teacher at UTS, suggested that it is the role of the human learner to keep problematizing the outputs of generative AI tools, so that critical thinking remains alive in the learning journey: “AI seems to make things more efficient. And so, our job is to keep complicating and disrupting and making things harder.” Applying this insight to assessment, Former UTS teacher and current Innovation and Technology teacher at Branksome Hall, Isabella Liu, felt that any assessment should be clear about what the student - as a human learner - can do that AI cannot do: “So, us [we?], as humans, can provide strategic direction, emotional intelligence, critical thinking, while the AI can help with the data processing, pattern recognition and repetitive tasks.” 

This kind of ‘human-centered’ framework of assessment would also pay attention to how students demonstrate their developing capabilities as critical thinkers, emotionally-intelligent team-players, and ethical decision-makers. Biology and AP Research teacher Alan Kraguljac noted that while AI could be used for more time-consuming tasks in AP research, students’ success would ultimately hinge upon how they speak to their process in completing that work: “So, if you use the source, you should be able to explain exactly, what, why, you used that source – all those things.” This was a feeling shared by UTS’s Director of AI Integration, who stressed that placing more emphasis on “competencies that build a student’s ability to formulate questions, sequence them logically, connect answers from various prompts, and read and fact-check the results” in AI-partnered assessment tasks would make them better-positioned for entering the careers market, as such skill sets are also “highly needed in industry.”

He also suggested replacing assessment tools such as tests and quizzes with portfolios, which are artifacts that can effectively demonstrate the evolution of a student’s learning journey. Such an approach shifts the area of focus from product to process. In addition to supporting critical thinking, a process-based success framework would also emphasize, with renewed vigour, the social-emotional aspects of learning. Director of the Eureka! Institute, Dr. Kimberley MacKinnon, referred to a talk held at UTS by Dr. Susan McCahan, University of Toronto Associate Vice-President and Vice-Provost, Digital Strategies, and Dr. Steven Katz, Associate Professor at the Ontario Institute for Studies in Education, in which they discussed ‘power skills’ such as listening and demonstrating empathy, interrupting the longstanding assumption that such ‘soft’ capabilities are ultimately less important than a student’s final performance on a culminating task. Natalie Cannistraro, Program Director of UTS’s Innovation Lab, argued that such ‘power skills’ could form part of the ‘human’ learning and work habits that teachers should now be assessing with enhanced interest and criticality. 

However, ultimately, any new approaches to assessment in response to GenAI must also be complemented by a curriculum that addresses how these tools function. This could, as Isabella suggested, occur by demystifying the ‘hallucination’ piece of AI or reckoning with its environmental impact. Any protocols on the use of GenAI in assessments and beyond must also account for the ethics of whether or how to use such technologies. 

Alongside (re)valuing critical thinking, ‘power skills,’ and ethical decision-making in assessment, another significant factor is reminding students that they themselves are vital contributors to knowledge and that assignments should account for this more explicitly. Perhaps rediscovering ‘our humanity’ as teachers and learners is a necessary move as we continue to navigate disruptive technologies in our schools and beyond? At Eureka! and the teachers with whom we work, we will continue to reflect upon our principles and practices as we prepare young people to thrive in the future. 


#utseureka