Why don’t you just make the AI-assisted parts yourselves?

We can, and in most cases we already know how.

We use AI because it can dramatically reduce the amount of repetitive, exploratory, or mechanical work between an idea and the finished result. It does not replace skills we do not have. It helps us use the skills we already have more efficiently.

For example, we may use it to quickly explore several visual directions before manually choosing and refining one; clean up or reconstruct parts of product mockups starting from our own photographs; remove distracting backgrounds; test different compositions; help reorganize or refine text we have already planned; compare wording; debug or review code; prototype an interface before implementing it properly; analyze technical problems; structure large amounts of information; or test different solutions before deciding which one is worth developing further.

Without AI, we could still do these things. We did many of them before these tools existed. They would simply take considerably longer.

This distinction matters to us. We do not type something like “make me a nice Game Boy cover” and publish whatever comes back. Our prompts are often detailed working documents containing the concept, composition, visual hierarchy, technical restrictions, references, dimensions, hardware limitations, things that must not appear, things that must remain unchanged, production requirements, and instructions developed through previous tests. A single project can involve many weeks of iterations, corrections, rejected outputs, manual edits, and new instructions before anything becomes usable.

The prompt itself is therefore part of our creative and technical work. Under Italian copyright law, however, length or complexity alone does not automatically make a prompt copyright-protected. The relevant law is Law No. 633 of 22 April 1941 on copyright, as amended in 2025 by Law No. 132/2025. Italian law protects creative works resulting from human intellectual work, including works made with the assistance of artificial intelligence. A sufficiently original prompt that expresses the author’s own creative choices may itself qualify for copyright protection, just like another written work, while a simple generic instruction may not.

The same principle applies more broadly to what comes out of the process. What matters is the human contribution. We define the concept, decide what problem has to be solved, establish the constraints, prepare or choose references, direct the iterations, reject unsuitable results, make corrections, combine elements where appropriate, perform the manual or technical work still required, and decide what becomes part of the final product.

AI is particularly useful to us because NOW IN TIME works across very different areas. A normal week can involve graphic design, product photography, Game Boy development, browser tools, 3D printing, copywriting, archival research, WordPress work, debugging, packaging, customer documentation, or adapting old technology to a completely new purpose. AI can move quickly between those contexts and help with the intermediate work, while the knowledge needed to judge whether its answer is actually good still has to come from us.

And that last part is important: AI can produce something plausible very quickly. Plausible is not the same as correct.

We regularly reject AI output because it is technically wrong, visually weak, historically inaccurate, physically impossible to manufacture, incompatible with the hardware involved, inconsistent with the rest of the project, or simply not good enough. Being able to recognize those failures requires the same knowledge we would need to do the work without AI in the first place.

So why use it?

For much the same reason a designer uses Photoshop instead of developing photographs in a darkroom, a programmer uses a debugger instead of reading every machine instruction by hand, or a 3D designer uses CAD instead of carving every prototype from a block of material. A tool can remove unnecessary labour without removing the human decisions that make the work ours.

AI gives us more time for the parts we actually care about: designing the object, solving the strange technical problem, testing it on real hardware, improving the details, making the physical product, and deciding whether the final result deserves the NOW IN TIME name.

For a detailed explanation of where and how we use AI, see “Do you use artificial intelligence in your products or website?”