[2025-03-25] New authenticity score replaces risk

Risk (low, medium, high) is replaced with the authenticity score. This new score shows an integer scale (typically 1-5) indicating how authentic the submitted work is. The score is used as the primary factor and evidence in the assignment view, showing the table of submissions. The score is calculated based on an aggregation of all the validation tests conducted (linguistic analysis and question sessions).

A cut-out of the submission table showing two submissions, one with a 2/5 authenticity score (blue) and one with a 4/5 authenticity score (green).

A cut-out of the submission table showing two submissions, one with a 2/5 authenticity score (blue) and one with a 4/5 authenticity score (green).

A high score generally indicates that the system validates that the submitter is the sole author of the submission.

Not all assignments should require the same authenticity score

The authenticity score must always be interpreted in context with the institution’s assignment policy and the guidelines set out in each individual assignment. If students are allowed to touch up their language using tools, such as Grammarly or ChatGPT, you might accept a lower authenticity score than if you expect students to write completely unassisted.

Similarly, when using text-matching software (plagiarism checkers), a 3% similarity score in a 100-page thesis is very different from a 20% similarity score in a 3-page essay. The expected authenticity score should therefore reflect the expectations for each assignment.

Updated scores for each validation test

The submission list table in the assignment view has been updated to show new scores for each validation test.

Text analysis includes the authorship verification step, which compares the writing style in the submission with the student’s linguistic signature. This column shows the percentage of text in the submissions the system could verify. The human text column shows the percentage of text in the submissions the system could validate as human-written.

Question scores show both free-text question scores and the close gap-match question scores. Each score indicates how many questions the student answered correctly, with partial corrects counting as 0.5.

The new authenticity score is calculated as the average of all four (4) scores, categorised into the 1-5 scale.

Each column is sortable by clicking on the column header.

A cut-out of the submission table in the assignment view, showing a submission with a 5/5 authenticity score, 100% verified text, 100% human text, 4.5/5 free-text score and a 5/5 cloze score.

A cut-out of the submission table in the assignment view, showing a submission with a 5/5 authenticity score, 100% verified text, 100% human text, 4.5/5 free-text score and a 5/5 cloze score.

Updated integrity report with the new authenticity scale

The integrity report has been updated to reflect the same scale shown in the submission table. The same values and scores that are available in the table are all displayed in the report.

An integrity report showing a 100% verified text, 100% human text, 4.5/5 free-text question score and a 4/4 cloze question score.

An integrity report showing a 100% verified text, 100% human text, 4.5/5 free-text question score and a 4/4 cloze question score.

The text analysis shows which parts of the text have been verified with different shades of green, corresponding to the confidence in the result.

Screenshot 2025-03-24 at 6.36.57 PM.png