AI Art and Portfolios: What Admissions Reviewers Look For Now

The Short Answer

Admissions reviewers typically look for two things when AI appears in a portfolio: honest disclosure of what was used and where, and clear evidence that the thinking belongs to the student. Schools increasingly accept AI as one tool inside a documented process. What they do not accept is substitution — generated images standing in for a student's own looking, making and judgement. Policies differ between schools and are still moving, so disclosure and evidence are the two safe constants.

The Disclosure Principle

Treat AI like any other medium: name it. A caption that states the tool and the role it played turns a question mark into ordinary documentation, the same way a printmaker names a technique or a photographer describes a process.

• State the tool and the stage it entered — ideation, reference, texture, iteration
• Show prompts and selected outputs as process material, the way thumbnails or contact sheets would appear
• Make clear what the student did before and after the generated step

Undisclosed use that surfaces later does more damage than the use itself, because schools typically treat it as an integrity question rather than a technique question.

Tool or Substitute: Where Reviewers Draw the Line

The distinction reviewers typically draw is between AI that extends a practice and AI that replaces one.

• Extends — generating variations to study, building moodboards, exploring texture and colourways, then making choices the student can explain
• Replaces — presenting generated finals as finished personal work, with no visible route from idea to outcome

There is a simple test a family can run at home: remove the AI-assisted pieces and look at what remains. A portfolio that still shows observation, development and a personal direction is using AI from a position of strength. If removing them leaves little behind, the portfolio is not ready yet — and that would be true whatever tool had filled the gap.

How Reviewers Typically Recognise It

Reviewers read portfolios against each other all day, and generated work tends to announce itself through mismatch — finished images far beyond the skill visible in the sketchbook, uniform surface polish with no development trail, or a style that changes from piece to piece without explanation.

Interviews do the rest of the work. Applicants are typically asked to talk through decisions — why this composition, what was rejected, what would you change next. Work a student genuinely made, with or without AI somewhere in the loop, survives those questions comfortably.

Policies Are Still Moving

There is no single rule across institutions. Some schools now publish AI guidance for applicants, some ask for a declaration of tools used and some restrict AI in specific submissions such as observational studies. Others have said little so far. Guidance typically gets revised from cycle to cycle as the tools change, which puts the responsibility on applicants to check.

• Read each school's current requirements in the cycle you apply — last year's advice may already be stale
• Where guidance is ambiguous, disclosing more than required is the safer reading
• Keep a process log for any AI-assisted piece so the record exists if a school asks for it

A Sensible Way to Work With It

For students building portfolios now, the workable position is straightforward. Build the foundations first — observational drawing, self-directed projects, a dated record of development. Let AI enter where it genuinely helps, document it as it happens and caption it without fuss. A portfolio built this way does not depend on any school's current policy, because it already shows what every policy is trying to protect: evidence of a student who can see, think and make.