"The Nature of Learning Goals Is Dynamic Rather than Fixed"
Mannheim (GER), April 2022 - Prof. Dr. Dirk Ifenthaler is involved in the development and evaluation of AI-based personalization approaches for in-company continuing education in a University of Mannheim project funded by Germany’s Ministry of Education and Research. The effort is called KAMAELEON, the German acronym for "Contact-Based, Adaptive Measures for Effective Learning Support in Continuing Education", and it is also the title of his presentation at the LEARNTEC congress, 01 June at 16:15.
The KAMAELEON effort addresses issues such as whether online continuing education programs can effectively address the participants’ individual learning needs and investigates the optimal ways to achieve individualization of the learning experience. Perhaps surprisingly, in Ifenthaler’s opinion, personalized and adaptive learning environments function according to the same principles as online shopping platforms and streaming portals!
What is the relationship between personalized and adaptive learning environments to academic success and learning in general?
Prof. Dr. Dirk Ifenthaler: Findings on personalized and adaptive learning environments are generally the results of research done in the context of higher education. In a meta-analysis recently published by my research group (Ifenthaler and Yau, 2020), factors in the relationship between personalized and adaptive learning environments and learning success were identified from over 6,000 studies.
The findings indicate that analytics-based approaches are speciously used as data driven methods to recognize risk situations associated with learning success. Nevertheless, only a few effective pedagogical intervention strategies related to learner achievement were actually identified. Furthermore, it was not possible to observe evidence of any widespread implementation of such systems. Consequently, there is a great need for research and implementation in this area at universities. In the field of vocational education, there are as yet only meager findings regarding the effectiveness and implementation of personalized and adaptive learning environments.
You have played a key role in the KAMAELEON project, working on further development of the individualization of learning opportunities, for about a year now. Are your findings at the university level applicable in vocational training?
Prof. Dr. Dirk Ifenthaler: Indirectly, our findings from university research can be integrated into the KAMAELEON project, whose goal - by means of the edyoucated learning platform - is to develop and evaluate AI-based personalization approaches for corporate continuing education. In this effort, two groups of factors, the learners’ learning environments and the dynamic knowledge acquisition goals in continuing education processes, will be scrutinized in greater detail.
In one of the next phases, the AI-based personalization approaches that are developed will be evaluated iteratively in model tests with several continuing education programs from various industries. In the process, conditions for sustainable implementation (i.e., critical success factors and hurdles) that are transferable will be identified.
Has KAMAELEON already produced new findings regarding the use of artificial intelligence?
Prof. Dr. Dirk Ifenthaler: Our analyses make clear that although digital media and technologies are increasingly being used in corporate continuing education, online continuing education is often only available as a static comprehensive package, thus emulating the "one-size-fits-all" strategy of traditional classroom training. Both personalized and adaptive learning strive to support the learning experience by adapting it to the learner's current situation.
In this process, whereas personalized learning places a focus on individual learner characteristics, adaptive learning focuses on the continuous and ongoing collection of information. However, in the context of AI-based technologies, the lines between personalized and adaptive learning are beginning to blur. AI-based technologies make comprehensive collection and analysis of both static individual characteristics as well as continuously changing information possible. Thus, they combine personalized and adaptive learning.
In a sense, personalized and adaptive learning environments function on the same principle as online shopping platforms and streaming portals. In the educational context, huge amounts of student data are collected in order to analyze learning behavior and automatically adapt the training programs to the learner’s individual situation.
The applications and functions of personalized and adaptive learning environments are multifaceted. For example, learning content and materials such as specialized exercises, videos or texts can be adapted to individual interests and preferences, as well as the learner’s current knowledge level. Personalized user interfaces, adaptive feedback, learning paths, and learning target recommendations adapted to the learner’s personal prerequisites and needs represent yet further functions of personalized and adaptive learning environments.
The support of informal workplace learning is another potential use scenario of personalized and adaptive learning environments. In the work context, learning extends beyond formal training opportunities, often taking place informally during the work itself. When employees work on tasks, solve problems, or interact with team members, they expand their knowledge and skills - often without being aware of it. When employees are confronted with a problem while working on a task, information provided by personalized and adaptive learning data can, for example, suggest in-house team members who have already encountered this issue and are possibly in a position to provide assistance. Personalized and adaptive learning environments can also encourage employees to reflect on their work performance and results and suggest recommendations for improvement.
To what extent can dynamic learning goals be supported by context-based learning support measures?
Prof. Dr. Dirk Ifenthaler: An assumption of the KAMAELEON project is that learning goals are the desired outcomes or target states for which learners strive in continuing education. Goal-setting theory asserts that goals have a positive effect on performance by focusing attention on relevant tasks and increasing energy and persistence. Empirical studies support this assumption and show that certain characteristics of learning goals (e.g., goal commitment, intrinsically defined vs. externally set goals) can have varying effects on learning success and performance. Therefore, in personalized and adaptive learning environments, information concerning learning goals should be considered in order to support learning processes and success.
In the work context, learning goals can be defined intrinsically (e.g., desired career or position change, personal development), as well as set externally (e.g., requirements by superiors, organizational demands). Furthermore, due to the constantly changing requirements of the job, learning objectives are not fixed: rather, they are dynamic in nature and can change during the course of the training process.
The KAMAELEON project’s findings so far demonstrate that conventional learning platforms are often not flexible enough to meet the requirements of individualized online professional development due to the heterogeneity of current knowledge levels, personally preferred learning strategies, and learners’ constantly changing learning goals. Fundamental problems in the use of AI in online continuing education include context dependency, as well as the fragmentation and bias of available data.