Table 2 Relevant properties of AI-based tools for use in teaching and learning.

From: Artificial Intelligence in studies—use of ChatGPT and AI-based tools among students in Germany

Category

Characteristics

Interaction

Adaptation of the interaction to the user (e.g. personalised adaptation to previous chat history)

Error detection/correction during input (e.g. grammar)

Error prevention during output (e.g. misinformation)

Reasoning

Degree of scientificity (e.g. citation/source citation)

Explainability of the decision (e.g. comprehensibility of the results)

Logical reasoning

Usability

Learning ability (e.g. AI learns from chat histories)

Communication with the system (e.g. colloquial vs. technical language)

User interface (e.g. clarity)

Possibilities of interactive input (e.g. input in text, language, image)

Possibilities of interactive output (e.g. output in text, language, image)

Size of the output (e.g. 10, 20, 50, 80 pages output)

Extent of input (e.g. 1, 2, 5, 10 pages of input)

Safety

Consideration of safety protocols (e.g. age restrictions)

Consideration of ethical standards (e.g. no discrimination)

Degree of data protection (e.g. handling of personal data)