Distinguish plagiarism, AI use and independent work
Plagiarism checks typically search for similarity with existing sources. AI detection attempts to classify statistical language patterns. These are different questions, and both approaches have limitations.
The rules of the school or university are decisive: What assistance is permitted, what must be disclosed and which work must be independent?
Why a detector score is not enough
- short or heavily edited texts are difficult to classify
- simple language may be incorrectly treated as suspicious
- different detectors can return different results for the same text
- a score does not identify the specific rule that may have been breached
A technical indication can prompt a conversation. It should not decide authorship or misconduct on its own.
A clean working process with AI
- Check applicable rules before beginning.
- Document the original question and outline.
- Read primary sources and cite them correctly.
- Record permitted AI assistance and tools.
- Verify every claim, number and reference.
- Be able to explain and defend the final work orally.
When uncertain, ask the teacher or examination office before submission.