Many colleges use AI-powered early warning systems (EWS) to provide support to students as soon asthey start their first semester. However, concerns arise regarding the fairness of an EWS algorithm whendeployed so early in a student’s college journey, especially at institutions with limited data collectioncapacity. To empirically address this fairness concern within this context, we developed a GPA predictionalgorithm for the first semester at an urban Japanese private university, relying exclusively on demographicand pre-college academic data commonly collected by many colleges at matriculation. Then weassessed the fairness of this prediction model between at-risk and lower-risk student groups. We also examinedwhether the use of 33 novel non-academic skill data points, collected within the first three weeksof matriculation, improves the model. Our analysis found that the model is less predictive for the at-riskgroup than their majority counterpart, and the addition of non-academic skill data slightly improved themodel’s predictive performance but did not make the model fairer. Our research underscores that an earlyadoption of EWS relying on pre-matriculation data alone may disadvantage at-risk students by potentiallyoverlooking those who genuinely require assistance.