Why Screenshots and AI-Generated Artwork Aren't Suitable for Print

It happens most weeks. Someone needs a banner for Saturday, they've got a logo, and the logo arrives as a screenshot from their phone. Or increasingly, they've had a go at making artwork themselves with an AI image tool, and it looks brilliant on the screen they made it on.

Then we open it in the studio and the problems start.

This isn't us being awkward. A screen and a printing press are two completely different pieces of technology, and a file that's perfect for one is often useless on the other. Here's exactly what goes wrong, so you can spot it before you're up against a deadline.

Screenshots are low resolution by design

Your phone screen doesn't need much detail to look sharp. It's small, it's backlit, and you hold it at arm's length. A screenshot captures exactly what's on that screen and nothing more.

Print is measured in dots per inch, and 300dpi is the standard for anything you'll hold in your hand. Do the maths on a typical screenshot and it gets uncomfortable fast. A 1080 x 1080 pixel image from an iPhone is 91mm square at 300dpi — smaller than a beer mat. Stretch it across an A5 flyer and you're down to about 150dpi. Put it on an A1 poster and you're looking at roughly 38dpi, which will print as a soft, mushy mess with visible blocks around every edge.

There's no way round this. You cannot add detail that was never captured. Upscaling software makes a guess at what the missing pixels should look like, and on text and logos that guess is usually wrong — you get soft, wobbly letterforms and a strange plastic sheen across flat colours.

Compression damage doesn't wash out

Screenshots are saved as JPEGs or compressed PNGs, and every time the file gets shared it can be recompressed again. WhatsApp is the worst offender — it will happily crush a 2MB image down to 200KB before it reaches us.

On screen you won't notice. In print, JPEG compression shows up as grey mottling around sharp edges, particularly black text on white and any hard-edged logo. On a business card at close reading distance it's very obvious, and once it's baked into the file there's no removing it.

Screen colour and press colour are different systems

Screens make colour with light — red, green and blue combined to produce white. Presses make colour with ink — cyan, magenta, yellow and black laid on paper, absorbing light rather than emitting it.

RGB covers a much wider range of colour than CMYK ink can physically reach. Bright electric blues, vivid oranges and neon greens simply don't exist in four-colour process printing. If you send an RGB screenshot, something has to convert it, and the punchy blue you approved on your monitor comes back a shade duller and slightly purple.

The fix is designing in CMYK from the start, or at minimum knowing which of your brand colours are going to shift so you're not surprised.

AI images have their own set of problems

AI art tools are genuinely useful for mood boards and rough concepts. As finished print artwork they cause more headaches than they solve.

The output is small. Most tools produce images at around 1024 x 1024 pixels. At 300dpi that's 87mm square. Fine for a sticker, nowhere near enough for a roller banner. Some newer tools go larger, but you're still working with a fixed pixels rather than artwork that scales.

It's not vector. Logos, text and line work need to be vector — mathematical shapes that stay razor sharp at any size, from a business card to a site board. AI tools output flat pixel images. There are no layers, no editable text, no separate logo element. If you want to move the phone number, you can't. We'd have to rebuild the whole thing.

Text comes out wrong. AI models still mangle lettering. You get almost-words, doubled letters, inconsistent kerning and characters that don't quite exist. At screen size your brain fills in the gaps. Blown up to A1 on a wall, it's the first thing everyone notices.

Small errors get enormous. Six-fingered hands, tools that don't make sense, patterns that don't line up, faint watermarks from the training data. Every one of these scales up with the artwork.

The rights are murky. Ownership and licensing of AI-generated images is still unsettled, and it varies by platform and by country. If you're putting an image on a shopfront, a vehicle or a thousand leaflets, you want to be certain you own it.

What to send us instead

• Vector files for logos and text. AI, EPS, SVG, or a PDF with fonts embedded or outlined. These are the gold standard and scale to any size.
• 300dpi at final print size. Not 300dpi at 50mm when you need 500mm.
• CMYK colour, with any brand colours specified as Pantone references if you need an exact match.
• 3mm bleed on every edge, with important text kept 5mm inside the trim.
• Original photographs at full size, straight from the camera or phone gallery — not screenshotted, not sent through WhatsApp.

Haven't got any of that?

Most people don't, and it's not a disaster. If the only version of your logo is a screenshot, we can redraw it as proper vector artwork — it's a one-off job and after that you own a file that works on everything, forever.

If you've used AI to generate a concept you like, bring it in. We'll happily use it as a starting point and build print-ready artwork around it. That's often the best use of these tools: getting an idea across quickly, then handing it to someone who can make it work on press.

Either way, send us the file before you're up against the wire. Five minutes of checking now saves a reprint later.

Got artwork you're not sure about? Email it to sales@cmyk-printing.co.uk, call the studio on 01282 421 000, or upload it here and we'll tell you honestly whether it'll print.