What Is Authentic Assessment?

A working definition#

Authentic assessment is a way of measuring learning through tasks that resemble real professional, civic or scholarly work. A nursing student interviews a patient and hands off the case. A business student negotiates a contract with a counterpart who has competing interests. An engineering student defends a design choice to a reviewer who has read the drawings. In each case the task is realistic, the student produces a performance or artifact, and an assessor judges it against criteria that were shared beforehand.

The word "authentic" describes the task, not the setting. A classroom task can be authentic if it asks for the same kind of thinking that the real situation demands. A workplace task can fail to be authentic if it only asks the student to repeat a procedure without understanding it.

Where the idea comes from#

Grant Wiggins made the case for authentic assessment in a 1990 ERIC Digest. His argument was that tests of isolated facts tell teachers little about whether students can use knowledge, and that assessment should engage students in worthwhile tasks that call for judgment and the kind of performance experts are expected to show. Later researchers turned the idea into frameworks that teachers can apply.

Gulikers, Bastiaens and Kirschner proposed five dimensions along which an assessment can be more or less authentic. Villarroel and colleagues later described authenticity in terms of realism, cognitive challenge and evaluative judgement. McArthur argued that the field should look past employability and ask how assessment contributes to students' wellbeing and to society.

The five dimensions#

The Gulikers framework is useful because it turns a vague quality into five questions you can ask about any assessment.

1. The task. Does it resemble what a practitioner would actually be asked to do?
2. The physical context. Does the setting, including tools and time, resemble the real one?
3. The social context. Does the student work alone, in a team or with a counterpart, as would happen in practice?
4. The result or form. Is the product something a practitioner would produce, and can others see and evaluate it?
5. The criteria. Are the standards the ones a professional community would use, and do students know them in advance?

An assessment does not need to score highly on all five. Most real courses cannot reproduce a hospital ward or a trading floor. The framework helps you decide which dimensions matter most for your learning outcomes and where you can reasonably compromise.

How it differs from a traditional exam#

Traditional exams remain useful. They are efficient, they are reliable for factual knowledge and they scale. The point is that an exam alone leaves a gap between knowing about a thing and being able to do it.

Examples across disciplines#

• Nursing. A student interviews a simulated patient, identifies the priority problem and gives an SBAR handoff to a colleague.
• Business. A student negotiates the terms of a supplier contract against a counterpart who has been briefed to resist.
• Teacher education. A student holds a conference with a simulated parent about a child's progress.
• Emergency management. A student briefs an incident commander on resource allocation and defends the priorities.
• Computer science. A student walks a reviewer through their code and explains why they chose one design over another.
• Humanities. A student defends a thesis argument against a reader's objections.

Several of these tasks are conversations. That is not an accident. Much professional work happens through speaking, and a conversation lets an assessor ask a follow-up question when an answer is thin. Our guide to what an AI oral exam is describes how that format works, and AI role-play in education covers the simulated-counterpart version.

Common misunderstandings#

Authentic does not mean unstructured. Good authentic assessments have explicit criteria and consistent scoring. Without them the grade depends on who marks the work.

Authentic does not mean it must occur in the workplace. A well-designed classroom task can capture the demands of real work. What matters is the thinking the task requires.

Authentic does not mean it cannot scale. It is true that observing every student perform a task costs time. Recorded presentations, structured oral exams and simulations with recorded transcripts all reduce the cost, though none removes it. Whether the trade-off is acceptable depends on the course.

Authentic does not mean AI-proof. A realistic written task can still be completed by a language model. What helps is asking students to explain and defend their work, which is covered in authentic assessment in the age of generative AI.

How to design an authentic assessment#

1. Start with the outcome. Write the thing a graduate should be able to do. If the outcome is "analyze a case," the task should require analysis.
2. Describe the real situation. Who faces this task in practice, what do they have in hand, and who judges the result?
3. Choose the form of evidence. A performance, a product, a conversation or a portfolio. Decide what record will exist for you to review.
4. Write criteria in the language of the field. Four to six criteria with clear performance levels are easier to apply consistently than a long list.
5. Share the criteria and an example. Students perform better when they know what good work looks like.
6. Plan for scale. Decide whether you will observe live, review recordings or sample. Add a practice attempt where possible.
7. Check validity. Ask whether a student could earn a high score without having the knowledge you care about. If yes, revise the task.

Grading authentic work#

Use an analytic rubric with criteria tied to the outcome, such as accuracy of reasoning, use of evidence, communication and professional judgment. Score each criterion separately and keep notes that point to specific moments in the work. If more than one person grades, calibrate on a few samples first. Students benefit when feedback names what they did and what to change.

Authentic assessment and generative AI#

Generative AI changes the conditions under which students complete many take-home tasks. An authentic task helps because it often requires context-specific decisions, but it does not stop a student from asking a model to draft the artifact. The practical response is to add evidence of the student's own reasoning, such as an oral defense, a process record or a live simulation. Our guides on assessing AI-assisted student work and on AI oral assessment compared with AI detection take that further.

Where Rocketproof.ai fits#

Rocketproof.ai is one platform built around this approach. It runs structured oral defenses of submitted work and AI role-play scenarios, and it keeps the transcript and rubric evidence for instructor review. The authentic assessment page describes the workflows. The ideas above apply whether or not you use any particular tool.