However, leveraging technology can significantly enhance the feedback process. This paper explores the potential of using technology to deliver effective feedback in educational settings. It discusses various technological tools and methods that can be employed to provide timely, personalized, and actionable feedback to students.
EPFL students from various academic programs and levels have been actively involved in evaluating personalized feedback in authentic educational settings. This collaborative effort has culminated in a comprehensive study that has been shared with the academic community as a preprint at the European Conference on Technology Enhanced Learning. The study underscores the importance of personalized feedback in the learning process and its potential to significantly enhance student engagement and academic performance.
The study, conducted by researchers at the University of Washington, found that participants were more likely to identify AI-generated feedback on coding projects, with a success rate of 75%, compared to a success rate of 58% for logical proof tasks. This suggests that the nature of the task influences the ability to distinguish between human and AI feedback. The researchers believe that the difference in success rates may be due to the distinct characteristics of coding and logical proof tasks.
The paper, published in the Journal of Applied Psychology, emphasizes the importance of constructive feedback in the workplace. It suggests that effective feedback should not only highlight areas of improvement but also acknowledge strengths and provide clear guidance for future actions. The research conducted by Nazaretsky and her team involved analyzing feedback mechanisms in various organizations and their impact on employee performance and motivation.
“We were too focused on the technology and not enough on the human aspect,” he said. “We need to build trust in AI systems by ensuring transparency, accountability, and fairness.” Nazaretsky emphasized the importance of involving stakeholders in the development and deployment of AI systems. “We need to engage with users, policymakers, and other relevant parties to understand their needs and concerns,” he said.
“We should never forget that we are dealing with people, not just numbers. For example, when we analyze data on patient outcomes, we should consider the individual stories behind the statistics.
