AI Scoring for Short-Answer and Essay Assessments, Too... "Teachers' Role Is Even More Important"

Seoul Metropolitan Office of Education Holds Elementary Teaching and Assessment Innovation Conference on Oct. 28
"Curiosity and Trial and Error Are Key to Learning in the AI Era"
"Teacher Judgment Essential Even in AI Scoring... Teachers' Work Will Also Change"

As the era of artificial intelligence (AI) dawns, the importance of question- and inquiry-based lessons is growing in elementary schools, and changes to assessment methods, including short-answer and essay-type assessments, are expected to make teachers' roles even more important. Against this backdrop, a conference on innovation in elementary teaching and assessment brought together elementary school teachers to discuss how lessons and assessment methods are changing in the AI era, as well as the resulting role of teachers.

According to the education sector on Oct. 29, the Seoul Metropolitan Office of Education held the "2026 Conference on Innovation in Elementary Teaching and Assessment" on Oct. 28 at Peace and Park Convention in Yongsan-gu, Seoul, for elementary school teachers and education professionals.

The event featured discussions on how students learn in the AI era, question-based lessons, and ways to use AI automated scoring for short-answer and essay-type assessments.

Seoul Metropolitan Office of Education Superintendent Jeong Geun-sik stressed the expansion of short-answer and essay-type assessments, saying, "Seoul is expanding short-answer and essay-type assessments, and 139 schools are preparing to implement them this year."

He also referred to the fact that the National Education Commission had planned to present discussions on the issue but postponed them, adding, "Schools are calmly preparing for this one step at a time."

Yonhap News Agency

Superintendent Jeong said it is important to help students develop the ability to think for themselves, but that putting this into practice in the classroom is not easy. "How to engage students with low motivation in lessons is also an important challenge," he said.

In a keynote address, Lee Ina, a professor in the Department of Brain and Cognitive Sciences at Seoul National University, explained the principles of human learning and the role of teachers under the theme "In the AI Era, How Do Students' Brains Learn and Remember?"

"The human brain learns through prediction errors, so the very concept of failure does not exist in the brain," Lee said. "We make predictions based on past experiences and learn when we encounter outcomes that differ from what we expected."

Lee Ina, a professor in the Department of Brain and Cognitive Sciences at Seoul National University, delivered the keynote address at the “2026 Conference on Innovation in Elementary School Teaching and Assessment” at Peace & Park Convention in Yongsan-gu, Seoul, on the 28th. Her presentation was titled “How Do Students’ Brains Learn and Remember in the Age of AI?” Photo by Oh Juyeon

She emphasized that "the process of students gaining firsthand experience and learning through trial and error is important." AI can provide knowledge quickly, she explained, but it cannot replace this process.

"Few things fuel the desire to learn as much as curiosity," Lee said. "In elementary school, it is very important to let children keep doing what they want to do and to create situations that make that possible."

In the sessions that followed, participants moved between discussions organized around themes such as "Changing Student Response Characteristics and Ways to Communicate," "Fostering Student Agency Through Student Questions," and "A Study on a Seoul-Style Lesson Model Based on Questions and Inquiry."

"Exploring Directions for Short-Answer and Essay-Type Assessment and AI Automated Scoring" was also discussed during the sessions.

Lee Jeonghun, a research fellow at the Korea Institute for Curriculum and Evaluation, who presented on the topic, said, "When the system was first introduced, the aim was to use large language models (LLMs) to reduce teachers' workloads, but the limitations of the system and LLMs mean that teacher involvement is unavoidable."

He added, "Having AI prepare a first draft can shorten grading time, but teachers need to exercise their judgment, and there must be a process for checking the supporting evidence."

At the “2026 Conference on Innovation in Elementary School Teaching and Assessment,” held on the 28th at Peace and Park Convention in Yongsan-gu, Seoul, Lee Jeonghun, a research fellow at the Korea Institute for Curriculum and Evaluation, delivered a presentation titled “Exploring the Future of Essay and Short-Answer Assessments and AI-Powered Automated Scoring.” Photo by Oh Juyeon

He cautioned against "automation bias," in which people accept AI-generated assessment results uncritically, and said teachers should continually review and improve AI scoring results and assessment criteria.

Regarding the event, Superintendent Jeong said, "We will work to realize in-depth learning centered on questions and inquiry, grounded in teachers' expertise and initiative, and ensure that innovation in teaching and assessment leads to the learning and growth of every student."

This content was produced with the assistance of AI translation services.