Visual Art & Design
2 gains2 Traces7 problems
Every measured gain and problem on the record for this topic, newest first.Too few to weigh yet. Updated .
One square per record, darker for stronger evidence:Evidence: 0 high6 moderate5 emerging
11 records so far, too few to weigh yet. Is your study missing? Add it
AI in visual art and design
A Trace is one record with both a gain and a problem, so it counts on both sides: 4 records show a gain and 9 show a problem. 11 context records are in the record list but not counted, because they report no measured gain or problem.
| Evidence | Gain side | Problem side |
|---|---|---|
| HighStrong study design, backed by at least two independent sources | 0 | 0 |
| ModerateA measured study or an official source | 3 | 4 |
| EmergingEarly, small, unreviewed or reported second-hand | 1 | 5 |
By type of evidence: 3 news reports; 2 interview studies; 2 controlled comparisons; 2 surveys; 1 observational study; 1 systematic review or meta-analysis. How evidence is scored
11 records so far, too few to weigh yet. A written summary appears once a topic has at least 20 counted records, with at least 5 on each side. Is your study missing? Add it
Across three recognition experiments with US adults, the ResMem neural network predicted which artworks would be remembered more accurately than laypeople or the artists who made them.
In three studies, crowdsourced participants could tell good AI art prompts from weak ones and wrote descriptive prompts for text-to-image generation.
Among more than 50,000 artists in an online art community, adopting text-to-image AI was linked to a 25% rise in creative productivity and a 50% rise in favorites per view.
A review of 38 publications found human artists fear displacement by AI image generators, while consumers and galleries criticize AI art.
In a survey of 426 AI art platform users, cultural bias in training data and algorithmic transparency were weighted lower than control over results and ease of customizing models, which the authors read as limited awareness.
In a survey of 318 Chinese fashion designers, those who saw more risk in AI-generated content reported lower autonomy, competence and relatedness, which in turn predicted lower intention to use it.
2 records
"BOLEX B8 MOVIE CAMERA 8mm" by glen edelson, CC BY 2.0.By November 2024, researchers tested prompt engineering as a creative skill for AI art in three studies with crowdsourced participants, asking them to judge prompt quality, write prompts, and refine them.
International Journal of Human–Computer Interaction
Published February 28 2026, this peer-reviewed study examined how art students negotiate creative identity when working with generative AI.
ace