Judging AI Output Like an Art Director
Develop a critical eye for generated images and video and give feedback that improves them
A taste of a lesson
My team says a campaign image 'feels AI' but nobody can say why. It's technically clean. How do I diagnose it?
Go through the usual tells one by one. Skin: poreless and evenly lit? Grade: teal shadows, orange highlights? Composition: centred subject, symmetrical background, shallow focus everywhere? Detail: everything equally sharp and glossy, with no imperfection or wear? Casting: faces that look like the model's average? Often it is three or four together. Then compare it with a strong human photo from your category at the same size. Which of those tells do you see in your image?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Evaluate generated work with a structured framework
- Spot generic AI aesthetics and category clichés
- Select efficiently from large sets in staged passes
- Write specific, actionable feedback and run team critiques
Lesson plan
- 1 A framework for judging Evaluate work in a consistent order, from brief to craft. Start
- 2 Seeing generic AI Recognise the visual tells and clichés of generated work. Start
- 3 Staged selection Choose from large sets quickly and fairly. Start
- 4 Feedback that works Write precise, actionable feedback linked to the brief. Start
- 5 Representation and ethics as quality Treat stereotypes and misleading realism as quality failures. Start
- 6 Team critique and protecting taste Run critiques and guard against fatigue and drift. Start
Try asking
About this tutor
An advanced tutor for designers, creative leads and experienced makers who produce or approve large volumes of AI generated work. When generation is cheap, judgment becomes the scarce skill. You will build a structured way to evaluate output: fit to brief, idea strength, composition and hierarchy, light and colour, craft and artefacts, brand fit, originality versus cliché, representation and ethics. You will practise selecting from large sets, writing precise feedback, running critiques with a team and avoiding the fatigue and taste drift that come from looking at endless generated images.
Reviews
4.7
3 ratingsSample
- Magnus O.Sample
The staged selection method cut our review meetings in half. Thumbnail pass first is ruthless and effective.
- Priscila A.Sample
Feedback lesson changed how I brief our generalists. Visible terms instead of vibes.
- Dong-hyun L.Sample
The reference wall idea is simple and I can feel it working against the sameness of AI imagery.
About the teacher
Art direction for AI work: briefs, taste, identity, clients, rights and disclosure
9 tutors 374 lessons taught Sample
I teach the judgment side of creative AI: deciding what good looks like, explaining it to others and handling the rights and client questions that come with generated work. My background is in art direction for campaigns and identity projects, where my job was to say yes, no or not yet to other people's images. I teach that same eye....
See Ren's profile and tutorsMore like this
Other tutors on the same or nearby topics.