How can artificial intelligence assist teachers with formative and summative assessment?

In 2022, the European Commission published a set of ethical guidelines for educators to clarify popular and widespread misconceptions about the use of AI and data in education. AI encourages us to rethink traditional teaching, learning and assessment practices, including the role of teachers. However, concerns also exist about AI’s potentially disruptive aspects, including its potential to replace students’ own work in tests.

Educational assessment has the potential to be positively enriched by AI. AI can support teachers in their administrative responsibilities and enable them to focus on the pupil-centred aspects of their work. UNESCO recommends that educational stakeholders be aware of AI assessment tools and solutions, and adjust curricula to integrate AI where the benefits clearly outweigh the risks.

AI’s evolving role in assessment

The 2017 NESET analytical report on assessment practices defines four generations of e-assessment strategies:

• 1-2: Computerised/adaptive testing
• 3: Continuous measurement
• 4: Embedded assessment, intelligent measurements

AI-powered assessment is at the core of this fourth generation. Embedded assessment combines formative and summative assessment in the learning process. 

Current and future e-assessment strategies.
Source: NESET report / IPTS (European Commission)

AI changing assessment practices

A 2021 systematic review shows that most AI research in this area focuses on formative assessment, e.g. giving immediate feedback or automatic grading. A 2020 study on using AI for both formative and summative assessment in a data science course showed that students using the formative assessment system for ongoing self-assessment reported higher grades and better understanding of analytical methods. Summative teacher assessments using the automated tool reduced time spent and increased accuracy compared to manual grading.