Understanding PII Redaction

PII stands for Personally Identifiable Information; it is information that can identify a specific person, either on its own (such as a full name or student ID) or when combined with other details. PII redaction in Juried Assessment is an AI-assisted 'anonymization layer' that detects personally identifiable information in learner artifacts and masks it from assessors, allowing scoring to focus on the quality of the work rather than the student's identity.

The platform detects PII and contextual information in the artifact content (including common places like headers/footers and other embedded content), then replaces/blocks that text via blackout-style redaction.⁠⁠​ The redaction targets identifiers that are unambiguous, FERPA-protected, or known to introduce scoring bias. The goal is to support blind review and reduce the risk of bias while still allowing assessors to apply professional judgment to the work itself.⁠⁠

Direct Personal Identifiers

Direct personal identifiers are always redacted.

Direct personal identifiers are the clearest type of PII: data elements that can directly identify a specific person without requiring much additional context. In a Juried Assessment context, these identifiers are always redacted because they can reveal the learner’s identity and introduce bias risk:

• Student name (first, last, preferred)

• Student ID number

• Email address (institutional or personal)

• Username/login ID

Even though these identifiers are not directly typed into the main content of an artifact, direct identifiers often appear in:

• Cover pages/title blocks (name, course, instructor)

• Headers/footers (name + page numbers)

• Document metadata (e.g., “Author” field in Office/PDF metadata)

• Comments/track changes (names of commenters/editors)

Institutional and Contextual Identifiers

Institutional and contextual identifiers are always redacted.

Institutional and contextual identifiers are details that may not uniquely identify a student by name or ID, but can still bias reviewers (or help someone infer identity) by revealing the academic or instructional context surrounding the work.⁠⁠​ In a Juried Assessment context, these identifiers are always redacted because even when a student’s name is removed, reviewers can be influenced (consciously or not) by signals like perceived rigor or an experience, Instructor reputation/prestige, or timing/context clues about when/where the work occurred:

• Course number

• Course section number

• Instructor name

The identifiers most commonly appear in:

• Cover pages (EDUC 5555 – Section 01”, “Dr. John Smith”, “College of Smith”)

• Running headers/footers (course + section, instructor name, institution logo/name)

• Screenshots pasted into submissions (LMS pages, Gradescope/Canvas screens, etc.)

• Auto-generated filenames

• Timestamp lines (Submitted 2026-02-03 11:42 PM)

Supported File Types

PII redaction is supported for artifact formats where redaction can be applied reliably and auditably. Formats in which identity is intrinsic to the evidence are intentionally excluded from blind, juried workflows.

PII redaction is not supported for the following artifact formats