“AI-generated” on this shop means an image produced by a generative model from a text or image prompt, then selected, art-directed and curated by a person — Burak Basci — before it becomes a product. It is not a fully automated, unsupervised pipeline, and not a scan or photograph of someone else’s existing artwork. The EU AI Act’s Article 50 labelling duty, in force since 2 August 2026, specifically targets deepfake-style content with a concrete reference to reality — a real, identifiable person, place or event — so most stylised, abstract or invented-scene AI art sits outside that particular trigger, narrower than many buyers assume. Separately, the big marketplaces have their own AI-art rules: Etsy requires sellers to disclose AI use in the listing description, and Society6’s artist help page lists “No AI-generated art, or stock photos” among its criteria for new artists. Print quality is a different question again, governed by pixel resolution, not by how the image was made.
What “AI-generated” means in this catalogue
Every piece starts as a text or image prompt fed to a generative model. The model produces candidate images; a person — Burak Basci — reviews them, picks which ones are worth developing, directs follow-up prompts or edits, and curates the final set that goes on to become posters, canvas prints, framed posters, desk mats, mugs or apparel. No image here is published straight from the model without that review step, and none is a scan, photograph or reproduction of an existing artwork by someone else. That distinction — curated and directed, not autonomous and unsupervised — is also the detail that matters for the labelling question below, since it describes what kind of AI output this catalogue actually is before asking which rules apply to it. See About Burak and the shop for the full process.
Does EU law require a disclosure label on this kind of art?
What Article 50 actually covers
The EU AI Act’s Article 50 content-transparency duty has applied since 2 August 2026, and it has two separate parts. AI system providers must make synthetic audio, image, video or text output machine-detectable. Deployers — whoever publishes the content — must disclose it specifically when it is a “deep fake”: content defined narrowly as AI-generated or manipulated material that resembles a real or plausible person, object, place or event closely enough to appear authentic, with a concrete reference to reality and potential for deception. Routine generative art does not automatically meet that bar. In practical terms, a shop like this one would sit in the deployer role — the party that would carry a disclosure duty if a specific piece qualified as a deepfake — which is exactly why it matters which of the two categories below a given image falls into.
Concretely: an abstract composition, an invented landscape, or a stylised portrait of nobody specific — like Desert Dunes at Night Canvas Print, an invented night landscape of dark hills under a starry sky that doesn’t depict any real identifiable person, place or event — generally sits outside the deepfake trigger, because it isn’t presented as a genuine depiction of something real. A realistic-looking AI image clearly meant to depict an actual real person, such as a named public figure, would fall inside that scope instead. The difference is what the image claims to show, not that it was AI-generated.
This is general information about a 2026 EU regulation, not legal advice, and rules can be updated — if this matters for your own use case, check current guidance.
Why this shop discloses anyway
Separately from that narrow legal trigger, the same consumer-transparency expectations that already apply to describing any product’s real characteristics accurately are a sensible reason to disclose AI origin anyway — a trust practice, not a box being ticked for a specific law. That’s why this shop states its AI-directed process openly rather than leaving it to a product description footnote; see About Burak and the shop for how each piece is made and curated. The reasoning is the same one behind describing any material, dimension or finish accurately on a product page — a buyer should be able to find out how something was made without having to ask.
How Etsy and Society6 treat AI art
Two large marketplaces have their own AI-art rules, independently of any EU law. Etsy’s Creativity Standards allow AI creations generated from a seller’s own prompts, and require sellers to disclose within the listing description if an item was created with the use of AI. Society6’s artist help page lists “No AI-generated art, or stock photos” among what it looks for in new artists, and says every application is reviewed individually. We have not checked how either platform enforces these rules, or when they were introduced.
Neither change is a verdict on whether AI-generated art is good or bad art. The honest takeaway is simpler: “AI-generated” by itself isn’t a quality signal one way or the other — it’s become something buyers increasingly expect to be told about plainly, which is exactly what disclosure, here and elsewhere, is for.
Print quality is a resolution question, not an AI question
The pixel math behind a sharp print
A print looks sharp or soft because of how many pixels the source file has relative to the size it’s printed at — nothing about whether a model or a camera produced those pixels changes that math. As the poster size guide works out, a 6 × 8 in image at 300 pixels per inch (ppi) is 1800 × 2400 pixels (6 × 300 = 1800, 8 × 300 = 2400). The same arithmetic applies to a larger print: a 9 × 11 in print at 300 ppi needs 9 × 300 = 2700 pixels by 11 × 300 = 3300 pixels, i.e. 2700 × 3300 pixels.
Why a large canvas needs fewer pixels per inch
Pixel density needed also depends on viewing distance, a standard point in print production, not something specific to AI art: a print viewed up close — a desk mat, a small poster on a desk — benefits from the full 300 ppi, because the eye is close enough to resolve that much detail. Move further back from the same surface and the eye simply can’t distinguish detail at that fine a scale any more, so adding pixels beyond a certain density stops making a visible difference. A large canvas is normally viewed from further back than a desk mat or small poster, so roughly 150 ppi already reads as sharp at a typical viewing distance for that format. Worked example: a 36 × 24 in canvas at 150 ppi needs 36 × 150 = 5400 pixels by 24 × 150 = 3600 pixels, i.e. 5400 × 3600 pixels — noticeably fewer total pixels than the close-viewing math above would suggest for a print of that size, precisely because it’s meant to be seen from farther away.
| Print size | Viewing distance | ppi used | Pixel dimensions (arithmetic) |
|---|---|---|---|
| 6 × 8 in | close (desk mat, small poster) | 300 | 6×300 = 1800, 8×300 = 2400 → 1800 × 2400 px |
| 9 × 11 in | close to medium | 300 | 9×300 = 2700, 11×300 = 3300 → 2700 × 3300 px |
| 36 × 24 in canvas | far (wall canvas) | 150 | 36×150 = 5400, 24×150 = 3600 → 5400 × 3600 px |
Upscaling can’t invent missing detail
Enlarging a small or low-detail file does not recreate detail that was never captured in the first place — software can smooth or sharpen existing pixels, but upscaling cannot invent information that isn’t there. That rule applies exactly the same way to an AI-generated source file as it does to a photograph: a 1200 × 1200 px AI image stretched to cover a 36 × 24 in canvas will look soft for the same reason an undersized photo would, not because it’s AI-generated.
The same resolution math applies when a customer uploads their own photo in the design studio rather than choosing a catalogue piece: the studio checks the uploaded file’s pixel dimensions against the chosen product and placement before the order goes through, for the same reason covered above.
Frequently asked questions
Does AI-generated wall art need a mandatory AI-disclosure label?
The EU AI Act’s Article 50 labelling duty, in force since 2 August 2026, specifically targets deepfake-style content with a concrete reference to a real person, place or event. Most abstract, landscape or invented-scene AI art falls outside that specific trigger. This is general information, not legal advice.
Is AI-generated art lower quality than a traditional photograph or illustration?
No. Print sharpness comes from the file’s pixel resolution at the size you print it, not from how the image was created. An AI-generated file and a photographed one follow the exact same resolution math.
How do Etsy and Society6 treat AI art?
Etsy allows AI creations based on a seller’s own prompts and requires sellers to disclose the use of AI in the listing description. Society6’s artist help page lists “No AI-generated art, or stock photos” among its criteria for new artists. Both point to a market expecting clear disclosure or human-made work, not a verdict on AI art’s quality.
Can I tell in advance if my chosen size will look sharp?
Yes. Check the pixel dimensions implied by the print size, shown with worked examples above and in the size guide. Larger formats like canvas need proportionally fewer pixels per inch than a close-viewed print, because they’re normally viewed from farther away.
For sizing and framing questions beyond this one, see the full buying guides hub.
To see the artwork itself, browse the posters and canvas prints collections; each product page states that the artwork is AI-generated.
An AI-assisted buying guide from benotable Art. Featured artworks are AI-generated; check each product page for current sizes and finishes.