CIFAR describes adaptive AI tools and changing classroom assessment
CIFAR has profiled University of Alberta researcher Carrie Demmans Epp’s work on adaptive learning software and questions schools face as students use generative AI. Demmans Epp says adaptive systems can help teachers see where individual students are struggling.
In a rural Alberta pilot, CIFAR reports preliminary improvement, but says formal findings are not yet published. The article also discusses language bias in automated essay grading and Demmans Epp’s view that assessment should value students’ learning process, not only finished work.
Adaptive practice is meant to give teachers a clearer view of learning
CIFAR describes Carrie Demmans Epp, a Canada CIFAR AI Chair at Amii and University of Alberta associate professor, studying educational software through the EdTeKLA Research Group. The adaptive systems described track students’ accuracy and speed, then adjust exercise difficulty, hints or topics.
Demmans Epp says the software is intended to support teachers, not replace core instruction, and may take the place of independent practice for 20 to 60 minutes a week. CIFAR also discusses students using generative AI for schoolwork. Demmans Epp proposes teaching productive use and assessing stages of work, such as drafts and brainstorming logs.
A rural Alberta pilot has preliminary results, not published findings
CIFAR reports that Demmans Epp led a two-month pilot in rural Alberta with young students. Researchers checked progress three times. The article says preliminary results showed measurable improvement for all students. Children described as having the lowest cognitive abilities had no summer slide, and their September assessment scores exceeded their June scores.
According to CIFAR, the school division expanded the program. CIFAR says formal findings have yet to be published. The article also reports concerns about essay-grading systems favouring Canadian and British English over African-American English.
The article leaves the strength and reach of the results open
CIFAR says the pilot’s formal findings have yet to be published. The article does not give student numbers, assessment measures or a comparison group. It also does not say whether the preliminary gains continued or would appear in other schools. The article reports language bias in large language model essay graders, but does not name the models or detail the evaluations.
Look for published findings and details on classroom use
CIFAR says formal findings from Demmans Epp’s rural Alberta pilot have yet to be published; its account gives no release date. The article says the school division expanded the program. Demmans Epp says parents and educators need to know that tools have been rigorously tested for bias, particularly against children whose home language shapes their writing.